(character) Word or text variable to be plotted.

Word embeddings from textEmbed for the words to be plotted (i.e., the aggregated word embeddings for the "words" variable).

Word embeddings from textEmbed for individual words (i.e., the decontextualized word embeddings).

(character) Character string describing type of measure to be computed. Default is "cosine" (see also "spearmen", "pearson" as well as measures from textDistance() (which here is computed as 1 - textDistance) including "euclidean", "maximum", "manhattan", "canberra", "binary" and "minkowski").

(character) Method to aggregate the word embeddings (default = "mean"; see also "min", "max" or "[CLS]").

(numeric) Option to select words that have at least occurred a specified number of times (default = 0); when creating the semantic similarity scores.

Tibble from the textPlot function.

Select words to significance test that have occurred at least min_freq_words_test (default = 1).

Number of words per dimension to plot with extreme Supervised Dimension Projection value (default = 10). (i.e., even if not significant; duplicates are removed).

Number of words to plot according to their frequency (default = 10). (i.e., even if not significant).

Number of words to plot that are in the middle in Supervised Dimension Projection score (default = 10). (i.e., even if not significant; duplicates are removed).

Color for all the titles (default: "#61605e").

Variable to be plotted on the x-axes (default: "central_semantic_similarity", could also select "n", "n_percent").

Title (default: " ").

Label on the x-axes (default: "Semantic Centrality").

Length of the x-axes (default: NULL, which uses c(min(word_data$central_semantic_similarity)-0.05, max(word_data$central_semantic_similarity)+0.05); change this by e.g., try c(-5, 5)).

Length of the y-axes (default: NULL, which uses c(-1, 1); change e.g., by trying c(-5, 5)).

Type of font (default: NULL).

(HTML color codes. type = character) Colors of the words selected as plot_n_word_extreme (minimum values), plot_n_words_middle, plot_n_word_extreme (maximum values) and plot_n_word_frequency; the default is c("#EAEAEA", "#85DB8E", "#398CF9", "#9e9d9d", respectively.

Vector with minimum and maximum font size (default: c(3, 8)).

Jitter height (default: .0).

Jitter width (default: .03).

Size of the points indicating the words' position (default: 0.5).

Transparency of the lines between each word and point (default: 0.1).

Size of the points not linked to a word (default is to not show the point; , i.e., 0).

Transparency of the points that are not linked to a word (default is to not show it; i.e., 0).

Title of the color legend (default: "SCP").

Label on the color legend (default: "x").

Position on the x coordinates of the color legend (default = 0.02).

Position on the y coordinates of the color legend (default = 0.05).

Height of the color legend (default = 0.15).

Width of the color legend (default = 0.15).

Font size of the title (default = 7).

Font size of the values in the legend (default = 2).

Set different seed (default = 1007).

(character) The text to be cleaned.

(boolean)

(boolean)

(boolean)

(boolean)

(boolean)

(boolean)

(boolean)

(boolean)

(boolean)

(boolean)

(boolean)

(boolean)

(boolean)

A tibble with character variables.

One or several character variables; if its a tibble or dataframe, all the character variables will be selected.

Boolean. If the input (words) is a tibble/dataframe with several character variables, a total variable is computed.

The unit entropy is measured in. The default is to used bits (i.e., log2; see also, "log", "log10"). If a total score for several varaibles is computed,the text columns are combined using the dplyr unite function. For more information about the entropy see the entropy package and specifically its entropy.plugin function.

Option to remove NAs when computing mean, median etc (see under return).

(character string) Locale Identifiers for example in US-English ('en_US') and Australian-English ('en_AU'); see help(about_locale) in the stringi package

Logical; if TRUE, lists other available Python/Conda environments.

Logical; if TRUE, scans for OMP-related shared libraries.

Logical; if TRUE, includes full sessionInfo() output.

List of word embeddings or a single tibble

Logical. If TRUE, the word embedding name or a custom name will be attached to the name of each dimension. If FALSE, the attached part of the name will be removed.

Optional character. If provided and dim_names = TRUE, this custom name will be attached after each column (e.g., dim1_name).

Word embeddings (from textEmbed()).

Word embeddings (from textEmbed()).

(character) Character string describing type of measure to be computed; default is "euclidean" (see also measures from stats:dist() including "maximum", "manhattan", "canberra", "binary" and "minkowski". It is also possible to use "cosine", which computes the cosine distance (i.e., 1 - cosine(x, y)).

(boolean; from base::scale) If center is TRUE then centering is done by subtracting the embedding mean (omitting NAs) of x from each of its dimension, and if center is FALSE, no centering is done.

(boolean; from base::scale) If scale is TRUE then scaling is done by dividing the (centered) embedding dimensions by the standard deviation of the embedding if center is TRUE, and the root mean square otherwise.

Word embeddings (from textEmbed()).

(character) Character string describing type of measure to be computed; default is "euclidean" (see also measures from stats:dist() including "maximum", "manhattan", "canberra", "binary" and "minkowski". It is also possible to use "cosine", which computes the cosine distance (i.e., 1 - cosine(x, y)).

(boolean; from base::scale) If center is TRUE then centering is done by subtracting the embedding mean (omitting NAs) of x from each of its dimension, and if center is FALSE, no centering is done.

(boolean; from base::scale) If scale is TRUE then scaling is done by dividing the (centered) embedding dimensions by the standard deviation of the embedding if center is TRUE, and the root mean square otherwise.

Word embeddings (from textEmbed()).

Word embedding from textEmbed (from only one text).

(character) Character string describing type of measure to be computed; default is "euclidean" (see also measures from stats:dist() including "maximum", "manhattan", "canberra", "binary" and "minkowski". It is also possible to use "cosine", which computes the cosine distance (i.e., 1 - cosine(x, y)).

(boolean; from base::scale) If center is TRUE then centering is done by subtracting the embedding mean (omitting NAs) of x from each of its dimension, and if center is FALSE, no centering is done.

(boolean; from base::scale) If scale is TRUE then scaling is done by dividing the (centered) embedding dimensions by the standard deviation of the embedding if center is TRUE, and the root mean square otherwise.

A word-frequency data frame from textTokenizeAndCount

A word-frequency data frame from textTokenizeAndCount

A character variable or a tibble/dataframe with at least one character variable.

Character string specifying pre-trained language model (default 'bert-base-uncased'). For full list of options see pretrained models at HuggingFace. For example use "bert-base-multilingual-cased", "openai-gpt", "gpt2", "ctrl", "transfo-xl-wt103", "xlnet-base-cased", "xlm-mlm-enfr-1024", "distilbert-base-cased", "roberta-base", or "xlm-roberta-base". Only load models that you trust from HuggingFace; loading a malicious model can execute arbitrary code on your computer).

(string or numeric) Specify the layers that should be extracted (default -2 which give the second to last layer). It is more efficient to only extract the layers that you need (e.g., 11). You can also extract several (e.g., 11:12), or all by setting this parameter to "all". Layer 0 is the decontextualized input layer (i.e., not comprising hidden states) and thus should normally not be used. These layers can then be aggregated in the textEmbedLayerAggregation function.

(boolean) If TRUE append the variable name after all variable-names in the output. (This differentiates between word embedding dimension names; e.g., Dim1_text_variable_name). see textDimName to change names back and forth.

(string) Aggregated layers of each token. Method to aggregate the contextualized layers (e.g., "mean", "min" or "max, which takes the minimum, maximum or mean, respectively, across each column; or "concatenate", which links together each word embedding layer to one long row.

(string) Method to carry out the aggregation among the word embeddings for the words/tokens, including "min", "max" and "mean" which takes the minimum, maximum or mean across each column; or "concatenate", which links together each layer of the word embedding to one long row (default = "mean"). If set to NULL, embeddings are not aggregated.

(string) Aggregates to the word type (i.e., the individual words) rather than texts. If set to "individually", then duplicate words are not aggregated, (i.e, the context of individual is preserved). (default = NULL).

(boolean) Whether to also keep token embeddings when using texts or word types aggregation.

Number of rows in each batch

(bolean) TRUE warns and removes non-ascii (using textFindNonASCII()).

Option to select word embeddings linked to specific tokens such as [CLS] and [SEP] for the context embeddings.

Option to deselect embeddings linked to specific tokens such as [CLS] and [SEP] for the context embeddings.

(boolean) Provide word embeddings of single words as input to the model (these embeddings are, e.g., used for plotting; default is to use ). If using this, then set single_context_embeddings to FALSE.

The maximum length (in number of tokens) for the inputs to the transformer model (default the value stored for the associated model).

(numeric) Maximum number of tokens in a string to handle before switching to embedding text sentence by sentence.

(boolean) If TRUE this will turn on tokenizer parallelism. Default FALSE.

Name of device to use: 'cpu', 'gpu', 'gpu:k' or 'mps'/'mps:k' for MacOS, where k is a specific device number such as 'mps:1'.

Set to TRUE if the accessed model is gated.

The token needed to access the gated model. Create a token from the ['Settings' page](https://huggingface.co/settings/tokens) of the Hugging Face website. An an environment variable HUGGINGFACE_TOKEN can be set to avoid the need to enter the token each time.

Set the logging level. Default: "warning". Options (ordered from less logging to more logging): critical, error, warning, info, debug

(boolean; experiments) If TRUE the text is split using the DLATK-method; this method appears better for longer texts (but it does not return token level word embeddings, nor word_types embeddings at this stage).

(boolean) use a model with custom code on the Huggingface Hub

settings from textEmbedRawLayers().

Layers returned by the textEmbedRawLayers function.

(character or numeric) The numbers of the layers to be aggregated (e.g., c(11:12) to aggregate the eleventh and twelfth). Note that layer 0 is the input embedding to the transformer, and should normally not be used. Selecting 'all' thus removes layer 0 (default = "all")

(character) Method to carry out the aggregation among the layers for each word/token, including "min", "max" and "mean" which takes the minimum, maximum or mean across each column; or "concatenate", which links together each layer of the word embedding to one long row (default = "concatenate").

(character) Method to carry out the aggregation among the word embeddings for the words/tokens, including "min", "max" and "mean" which takes the minimum, maximum or mean across each column; or "concatenate", which links together each layer of the word embedding to one long row (default = "mean").

(boolean) If TRUE, provide the tokens used in the specified transformer model (default = FALSE).

(character) Option to only select embeddings linked to specific tokens in the textEmbedLayerAggregation() phase such as "[CLS]" and "[SEP]" (default NULL).

(character) Option to deselect embeddings linked to specific tokens in the textEmbedLayerAggregation() phase such as "[CLS]" and "[SEP]" (default NULL).

A character variable or a tibble with at least one character variable.

(character) Character string specifying pre-trained language model (default = 'bert-base-uncased'). For full list of options see pretrained models at HuggingFace. For example use "bert-base-multilingual-cased", "openai-gpt", "gpt2", "ctrl", "transfo-xl-wt103", "xlnet-base-cased", "xlm-mlm-enfr-1024", "distilbert-base-cased", "roberta-base", or "xlm-roberta-base". Only load models that you trust from HuggingFace; loading a malicious model can execute arbitrary code on your computer).

(character or numeric) Specify the layers that should be extracted (default -2, which give the second to last layer). It is more efficient to only extract the layers that you need (e.g., 11). You can also extract several (e.g., 11:12), or all by setting this parameter to "all". Layer 0 is the decontextualized input layer (i.e., not comprising hidden states) and thus should normally not be used. These layers can then be aggregated in the textEmbedLayerAggregation function.

(boolean) If TRUE, provide the tokens used in the specified transformer model. (default = TRUE)

(boolean) Wether to provide embeddings for each word/token type. (default = FALSE)

(boolean) Wether to dectonextualise embeddings (i.e., embedding one word at a time). (default = TRUE)

(boolean) Whether to keep token level embeddings in the output (when using word_types aggregation). (default= TRUE)

(character) Name of device to use: 'cpu', 'gpu', 'gpu:k' or 'mps'/'mps:k' for MacOS, where k is a specific device number. (default = "cpu")

(boolean) If TRUE this will turn on tokenizer parallelism. (default = FALSE).

The maximum length (in number of tokens) for the inputs to the transformer model (default the value stored for the associated model).

(numeric) Maximum number of tokens in a string to handle before switching to embedding text sentence by sentence. (default= 4)

Set to TRUE if the accessed model is gated.

The token needed to access the gated model. Create a token from the ['Settings' page](https://huggingface.co/settings/tokens) of the Hugging Face website. An an environment variable HUGGINGFACE_TOKEN can be set to avoid the need to enter the token each time.

use a model with custom code on the Huggingface Hub

(character) Set the logging level. (default ="error") Options (ordered from less logging to more logging): critical, error, warning, info, debug

(boolean) If TRUE sort the output to tidy format. (default = TRUE)

(list) Embedding(s) - including, tokens, texts and/or word_types.

(numeric) Number of dimensions to reduce to.

(string or matrix) Name or URL to scalar for standardizing the embeddings. If a URL, the function first examines whether it has been downloaded before. The string should be to a csv file containing a matrix with the pca weights for matrix multiplication. For more information see reference below.

(string or matrix) Name or URL to pca weights. If a URL, the function first examines whether it has been downlaoded before. The string should be to a csv file containing a matrix. For more information see reference below.

dataframe that at least contains one character column.

decontextualized/static space with a column called "words" and the semantic representations are in columns called Dim1, Dim2 (or V1, V2, ...) and so on (from textSpace, which is not included in the current text package).

default "null"; option to use either the "tk" of "df" space (if using textSpace, which has not been implemented yet).

method to aggregate semantic representation when their are more than a single word. (default is "mean"; see also "min" and "max", "concatenate" and "normalize")

Boolean, if TRUE append the variable name after all variable-names in the output. (This differentiates between word embedding dimension names; e.g., Dim1_text_variable_name)

(boolean) Lower case input.

(string) the language that was used for prediction/assessment/classification.

(numeric) the variable used for training (y).

(numeric) the outcome from the model (i.e., y_hat).

(string) If you are plotting errors between predicted and targeted scores, you can set the type to "prediction_errors", to produce two extra plots: distribution of scores and absolute error.

(integer) the n tile to split the data in (to show the most extreme tiles in different colours).

(integer) the number of language examples to show/select in each quadrant. When providing both x_variable and y_variable, each example is categorized into one of nine bivariate quadrants based on its position in the scatterplot (e.g., low–low, high–high, center). Within each quadrant, the function selects the most extreme examples by computing the distance to that quadrant’s corner: Corner quadrants (1, 3, 7, 9): Examples closest to the corner points (e.g., min x & max y) are selected using Euclidean distance. Edge quadrants (2, 4, 6, 8): Examples furthest along the relevant axis (x or y) are selected. Center quadrant (5): Examples closest to the mean of both x and y are selected.

(integer) the percentage of jitter to add to the data for the scatter plot.

(character vector) words required to be included in examples.

(string)

(string) = "darkblue",

= (string) "darkred",

(string) colors of the distribution plot

(string) file format of the figures.

(integer) The size of highlighted dots in the scatter legend.

(integer) The size of background dots in the scatter legend.

(numeric) The transparency alphe level of the dots.

(numeric) The transparency alphe level of the background dots. For example: c(1,0,1) result in one dot in each quadrant except for the middle quadrant.

(boolean) If TRUE, the estimate values are shown on the distribution plot axes.

(string) If a colour string is added, a regression line will be plotted.

(numeric vector) range of x axis (e.g., c(1, 100)).

(numeric vector) range of y axis (e.g., c(1, 100)).

(string) x-axis label of the grid topic plot.

(string) y-axis label of the grid topic plot.

(string)

(integer)

(string)

(string)

(integer)

(integer) The seed to set for reproducibility.

A character variable or a tibble including character variables.

A dataframe, where the first column contain text data, and the second column the to-be-predicted variable (numeric or categorical).

(string) Path to foundation/pretrained model or model identifier from huggingface.co/models

(string) Path to the output directory.

(Numeric) Proportion of the text_data to be used for validation.

(Numeric) Proportion of the text_data to be used for evaluation.

(String) Pretrained config name or path if not the same as model_name.

(String) Pretrained tokenizer name or path if not the same as model_name

(Numeric) The maximum total input sequence length after tokenization. Sequences longer than this will be truncated, sequences shorter will be padded.

(String or IntervalStrategy) — The evaluation strategy to adopt during training. Possible values are: "no": No evaluation is done during training. "steps": Evaluation is done (and logged) every eval_steps. "epoch": Evaluation is done at the end of each epoch.

(Integer) Number of predictions steps to accumulate the output tensors for, before moving the results to the CPU. If left unset, the whole predictions are accumulated on GPU/TPU before being moved to the CPU (faster but requires more memory).

(Numeric) Total number of training epochs to perform (if not an integer, will perform the decimal part percents of the last epoch before stopping training).

(Numeric, defaults to -1) Some models like TransformerXL or XLNet can make use of the past hidden states for their predictions. If this argument is set to a positive int, the Trainer will use the corresponding output (usually index 2) as the past state and feed it to the model at the next training step under the keyword argument mems.

(Numeric) Set the seed

Parameters related to the fine tuning, which can be seen in the text-package file inst/python/arg2.json.

A dataframe, where the first column contain text data, and the second column the to-be-predicted variable (numeric or categorical).

(string) Path to foundation/pretrained model or model identifier from huggingface.co/models

(string) Path to the output directory.

(Numeric) Proportion of the text_outcome_data to be used for validation.

(Numeric) Proportion of the text_outcome_data to be used for evaluation.

(Boolean) TRUE for regression tasks, FALSE for classification.

(String) Pretrained config name or path if not the same as model_name.

(String) Pretrained tokenizer name or path if not the same as model_name

(Numeric) The maximum total input sequence length after tokenization. Sequences longer than this will be truncated, sequences shorter will be padded.

(String or IntervalStrategy) — The evaluation strategy to adopt during training. Possible values are: "no": No evaluation is done during training. "steps": Evaluation is done (and logged) every eval_steps. "epoch": Evaluation is done at the end of each epoch.

(Integer) Number of predictions steps to accumulate the output tensors for, before moving the results to the CPU. If left unset, the whole predictions are accumulated on GPU/TPU before being moved to the CPU (faster but requires more memory).

(Numeric) Total number of training epochs to perform (if not an integer, will perform the decimal part percents of the last epoch before stopping training).

(Numeric, defaults to -1) Some models like TransformerXL or XLNet can make use of the past hidden states for their predictions. If this argument is set to a positive int, the Trainer will use the corresponding output (usually index 2) as the past state and feed it to the model at the next training step under the keyword argument mems.

(Numeric) Set the seed

label name in case of classification; e.g., label_names = c("female", "male").

Set to TRUE to solve error RuntimeError: MPS backend out of memory.Use PYTORCH_MPS_HIGH_WATERMARK_RATIO=0.0 to disable upper limit for memory allocations (may cause system failure). Monitor System Resources: If you decide to adjust this setting, closely monitor your system's resource usage to ensure it does not become unstable.

(boolean) If TRUE this will turn on tokenizer parallelism. Default FALSE.

Parameters related to the fine tuning, which can be seen in the text-package file inst/python/arg2.json.

(string) A variable or a tibble/dataframe with at least one character variable.

(string) Specification of a pre-trained language model that have been trained with an autoregressive language modeling objective, which includes the uni-directional models (e.g., gpt2).

(string) Device to use: 'cpu', 'gpu', or 'gpu:k' where k is a specific device number

(boolean) If TRUE this will turn on tokenizer parallelism.

(Integer) The maximum length the generated tokens can have. Corresponds to the length of the input prompt + 'max_new_tokens'. Its effect is overridden by 'max_new_tokens', if also set. Defaults to NULL.

(Integer) The maximum numbers of tokens to generate, ignoring the number of tokens in the prompt. The default value is 20.

(Integer) The minimum length of the sequence to be generated. Corresponds to the length of the input prompt + 'min_new_tokens'. Its effect is overridden by 'min_new_tokens', if also set. The default value is 0.

(Integer) The minimum numbers of tokens to generate, ignoring the number of tokens in the prompt. Default is NULL.

(string) Set the logging level. Options (ordered from less logging to more logging): critical, error, warning, info, debug

(boolean) Stop returning some incorrectly formatted/structured results. This setting does CANOT evaluate the actual results (whether or not they make sense, exist, etc.). All it does is to ensure the returned results are formatted correctly (e.g., does the question-answering dictionary contain the key "answer", is sentiments from textClassify containing the labels "positive" and "negative").

(boolean) Whether or not the output should include the prediction tensors (as token indices).

(boolean) If FALSE only the added text is returned, otherwise the full text is returned. (This setting is only meaningful if return_text is set to TRUE)

(boolean) Option to clean up the potential extra spaces in the returned text.

(string) Option to add a prefix to prompt.

By default, this function does not handle long generation (those that exceed the model maximum length).

(Integer) Set seed. (more info :https://github.com/huggingface/transformers/issues/14033#issuecomment-948385227). This setting provides some ways to work around the problem: None: default way, where no particular strategy is applied. "hole": Truncates left of input, and leaves a gap that is wide enough to let generation happen. (this might truncate a lot of the prompt and not suitable when generation exceed the model capacity)

(string) Select which columns to retrieve e.g., c("Name", "Path")

(string) Select which constructs concepts and/or behaviors to retrieve.

(boolean) TRUE downloads a new copy of the LBAM file

(string) The name of the model to know the number of layers of.

Set to TRUE if the accessed model is gated.

The token needed to access the gated model. Create a token from the ['Settings' page](https://huggingface.co/settings/tokens) of the Hugging Face website. An an environment variable HUGGINGFACE_TOKEN can be set to avoid the need to enter the token each time.

use a model with custom code on the Huggingface Hub

(string) The name of the model to be deleted.

(string) A variable or a tibble/dataframe with at least one character variable.

(string) Specification of a pre-trained language model for token classification that have been fine-tuned on a NER task (e.g., see "dslim/bert-base-NER"). Use for predicting the classes of tokens in a sequence: person, organisation, location or miscellaneous).

(string) Device to use: 'cpu', 'gpu', or 'gpu:k' where k is a specific device number

(boolean) If TRUE this will turn on tokenizer parallelism.

(string) Set the logging level. Options (ordered from less logging to more logging): critical, error, warning, info, debug

(boolean) Stop returning some incorrectly formatted/structured results. This setting does CANOT evaluate the actual results (whether or not they make sense, exist, etc.). All it does is to ensure the returned results are formatted correctly (e.g., does the question-answering dictionary contain the key "answer", is sentiments from textClassify containing the labels "positive" and "negative").

(Integer) Set seed.

Word or text variable to be plotted.

Word embeddings from textEmbed for individual words (i.e., decontextualized embeddings).

Lower case words

Set different seed.

Dataframe from textPCA

Select words to significance test that have occurred at least min_freq_words_test (default = 1).

Number of words that are extreme on Supervised Dimension Projection per dimension. (i.e., even if not significant; per dimensions, where duplicates are removed).

Number of words based on being most frequent. (i.e., even if not significant).

Number of words plotted that are in the middle in Supervised Dimension Projection score (i.e., even if not significant; per dimensions, where duplicates are removed).

Color for all the titles (default: "#61605e")

Title (default " ")

Label on the x-axes.

Label on the y-axes.

Manually set the length of the x-axes (default = NULL, which uses ggplot2::scale_x_continuous(limits = scale_x_axes_lim); change e.g., by trying c(-5, 5)).

Manually set the length of the y-axes (default = NULL; which uses ggplot2::scale_y_continuous(limits = scale_y_axes_lim); change e.g., by trying c(-5, 5)).

Font type (default: NULL).

The different colors of the words (default: c("#398CF9", "#60A1F7", "#5dc688", "#e07f6a", "#EAEAEA", "#40DD52", "#FF0000", "#EA7467", "#85DB8E")).

Vector with minimum and maximum font size (default: c(3, 8)).

Jitter height (default: .0).

Jitter width (default: .03).

Size of the points indicating the words' position (default: 0.5).

Transparency of the lines between each word and point (default: 0.1).

Size of the points not linked with a words (default is to not show it, i.e., 0).

Transparency of the points not linked with a words (default is to not show it, i.e., 0).

Title on the color legend (default: "(PCA)".

Label on the color legend (default: "(x)".

Label on the color legend (default: "(y)".

Position on the x coordinates of the color legend (default: 0.02).

Position on the y coordinates of the color legend (default: 0.05).

Height of the color legend (default 0.15).

Width of the color legend (default 0.15).

Font size (default: 7).

Font size of the values in the legend (default: 2).

Set different seed.

Dataframe from textProjection.

Select the k most frequent words to significance test (k = sqrt(100*N); N = number of participant responses) (default = TRUE).

Select words to significance test that have occurred at least min_freq_words_test (default = 1).

Select words to plot that has occurred at least min_freq_words_plot times (default = 1).

Select number of significant words in each square of the figure to plot. The significant words, in each square is selected according to most frequent words (default = 3).

Number of significant words to plot on each (positive and negative) side of the x-axes and y-axes, (where duplicates are removed); selects first according to lowest p-value and then according to frequency (default = 5). Hence, on a two dimensional plot it is possible that plot_n_words_p = 1 yield 4 words.

Number of words that are extreme on Supervised Dimension Projection per dimension. (i.e., even if not significant; per dimension, where duplicates are removed).

Number of words that are extreme in both x and y dimensions, considering overall distance from the origin in the Supervised Dimension Projection space. This selects words based on their combined extremity score, calculated as the Euclidean distance from (0,0). Ensures balance across all nine squares by selecting at least one extreme word per square if available.

Number of words based on being most frequent (default = 5). (i.e., even if not significant).

Number of words plotted that are in the middle in Supervised Dimension Projection score (default = 5). (i.e., even if not significant; per dimensions, where duplicates are removed).

(numeric) select random words to plot.

Color for all the titles (default: "#61605e").

(boolean) If TRUE, also plotting on the y-axes (default = FALSE, i.e, a 1-dimensional plot is generated). Also plotting on y-axes produces a two dimension 2-dimensional plot, but the textProjection function has to have had a variable on the y-axes.

Alpha (default = .05).

(boolean) Allow overlapping (TRUE) or disallow (FALSE) (default = TRUE).

(character) Method to adjust/correct p-values for multiple comparisons (default = "none"; see also "holm", "hochberg", "hommel", "bonferroni", "BH", "BY", "fdr").

(character) Metric to plot according to; "dot_product" or "cohens_d".

Title (default " ").

(character) Label on the x-axes (default = "Supervised Dimension Projection (SDP)").

(character) Label on the y-axes (default = "Supervised Dimension Projection (SDP)").

Manually set the length of the x-axes (default = NULL, which uses ggplot2::scale_x_continuous(limits = scale_x_axes_lim); change e.g., by trying c(-5, 5)).

Manually set the length of the y-axes (default = NULL; which uses ggplot2::scale_y_continuous(limits = scale_y_axes_lim); change e.g., by trying c(-5, 5)).

Font type (default = NULL).

(HTML color codes. Type = character) The different colors of the words. Note that, at the moment, two squares should not have the exact same colour-code because the numbers within the squares of the legend will then be aggregated (and show the same, incorrect value). (default: c("#398CF9", "#60A1F7", "#5dc688", "#e07f6a", "#EAEAEA", "#40DD52", "#FF0000", "#EA7467", "#85DB8E")).

Vector with minimum and maximum font size (default: c(3, 8)).

Jitter height (default: .0).

Jitter width (default: .03).

Size of the points indicating the words' position (default: 0.5).

Transparency of the lines between each word and point (default: 0.1).

Size of the points not linked with a words (default is to not show it, i.e., 0).

Transparency of the points not linked with a words (default is to not show it, i.e., 0).

Title on the color legend (default: "SDP").

Label on the color legend (default: "x").

Label on the color legend (default: "y").

Position on the x coordinates of the color legend (default: 0.02).

Position on the y coordinates of the color legend (default: 0.05).

Height of the color legend (default 0.15).

Width of the color legend (default 0.15).

Font size (default: 7).

Font size of the values in the legend (default: 2).

(string) Colour of the numbers in the box legend.

(boolean) Shows a point representing the aggregated word embedding for group 1 (default = FALSE).

(boolean) Shows a point representing the aggregated word embedding for group 2 (default = FALSE).

(boolean) Shows a point representing the aggregated direction embedding (default = FALSE).

Size of the points representing the group_embeddings1, group_embeddings2 and projection_embedding (default = 0.8).

Shape type of the points representing the group_embeddings1, group_embeddings2 and projection_embedding (default = 8).

Color (default = "black").

Color (default = "black").

Color (default = "blue").

(numeric) Set different seed (default = 1005)..

Explore where specific words are positioned in the embedding space. For example, c("happy content", "sad down") (default = NULL).

Specify the color(s) of the words being explored. For example c("#ad42f5", "green") (default = "#ad42f5").

Specify the names of the point for the aggregated word embeddings of all the explored words (default = "ALL_1").

Specify how to aggregate the word embeddings of the explored words (default = "mean").

Manually remove words from the plot (which is done just before the words are plotted so that the remove_words are part of previous counts/analyses) (default = NULL).

Set color to words that have higher frequency (N) on the other opposite side of its dot product projection (default = NULL).

Remove words that have higher frequency (N) on the other opposite side of its dot product projection (default = FALSE).

Provide a semantic space if using static embeddings and wanting to explore words (default = NULL).

Scaling word embeddings before aggregation (default = FALSE).

Settings for textOwnWordsProjection().

(character or r-object) model_info has four options, including: 1: An R model (e.g, saved output from one of the textTrain() functions). 2: The name specified in the L-BAM Documentation. For the following settings, remember to also set the model_type parameter: 3: Link to a text-trained model online (either in a github repo (e.g, "https://github.com/CarlViggo/pretrained_swls_model/raw/main/trained_github_model_logistic.RDS"or OSF https://osf.io/8fp7v) 4: Name or link to a fine-tuned model from Huggingface (e.g., "distilbert-base-uncased-finetuned-sst-2-english"). 5: Path to a model stored locally (e.g, "path/to/your/model/model_name.rds").

(character) Text to predict. If this argument is specified, then arguments "word_embeddings" and "premade embeddings" cannot be defined (default = NULL).

(character) Specify how the function should handle the model argument. The default is "detect" where the fucntion ttried to detect it automatically. Setting it to "fine-tuned" or "text-trained" will apply their respective default behaviors, while setting it to "implicit motives" will trigger specific steps tailored to these models.

(boolean) Updating the L-BAM file by automatically downloading it from Google Sheet.

(tibble; only for "text-trained"-model_type) Embeddings from e.g., textEmbed(). If you're using a pre-trained model, then texts and embeddings cannot be submitted simultaneously (default = NULL).

(tibble; only for "text-trained"-model_type) Variables to be appended with the word embeddings (x).

(boolean; only for "text-trained" models) If TRUE, x_appened is added before word embeddings.

(boolean; only for "text-trained"-models) Account for specific dimension names from textEmbed() (rather than generic names including Dim1, Dim2 etc.). If FALSE the models need to have been trained on word embeddings created with dim_names FALSE, so that embeddings were only called Dim1, Dim2 etc.

(boolean) If 'TRUE', the function will check whether the word embeddings (model type and layer) match the requirement of the trained model - if a mis-match is found the function till stop. If 'FALSE', the function will not verify.

(character column; only for "text-trained" models) If you provide the raw language data used for making the embeddings used for assessment, the language distribution (i.e., a word and frequency table) will be compared with saved one in the model object (if one exists). This enables calculating similarity scores.

(string or numeric; only for "text-trained" models) Default is to use the removal threshold used when creating the distribution in the in the training set ("trained_distribution_min_words"). You can set it yourself with a numeric value.

(boolean; only for "text-trained"-models) The model will by default be saved in your work-directory (default = TRUE). If the model already exists in your work-directory, it will automatically be loaded from there.

(numeric; only for "text-trained"-models) Determine threshold if you are using a logistic model (default = 0.5).

(boolean; only for "implicit-motives"-models) Show texts together with predictions (default = FALSE).

Name of device to use: 'cpu', 'gpu', 'gpu:k' or 'mps'/'mps:k' for MacOS, where k is a specific device number such as 'mps:1'.

(list; only for "implicit-motives"-models) Vector of participant-ids. Specify this for getting person level scores (i.e., summed sentence probabilities to the person level corrected for word count). (default = NULL)

(boolean; only for "text-trained"-models) If set to TRUE, embeddings will be saved with a unique identifier, and will be automatically opened next time textPredict is run with the same text. (default = TRUE)

(character; only for "text-trained"-models) Directory to save embeddings. (default = "wd" (i.e, work-directory))

(character; only for "text-trained"-models) Name of the saved embeddings (will be combined with a unique identifier). (default = ""). Obs: If no save_name is provided, and model_info is a character, then save_name will be set to model_info.

(vector; only for "implicit-motives"-models) Vector of story-ids. Specify this to get story level scores (i.e., summed sentence probabilities corrected for word count). When there is both story_id and participant_id indicated, the function returns a list including both story level and person level prediction corrected for word count. (default = NULL)

(R-object, tibble; only for "implicit-motives"-models) Insert your data here to integrate predictions to your dataset, (default = NULL).

(boolean; only for "implicit-motives"-models) If set to TRUE, word-embeddings will be averaged over the current and previous sentence per story-id. For this, both participant-id and story-id must be specified.

(boolean; only for "fine-tuned"-models) If TRUE this will turn on tokenizer parallelism.

(string; only for "fine-tuned"-models) Set the logging level. Options (ordered from less logging to more logging): critical, error, warning, info, debug

(boolean; only for "fine-tuned"-models) Stop returning some incorrectly formatted/structured results. This setting does CANOT evaluate the actual results (whether or not they make sense, exist, etc.). All it does is to ensure the returned results are formatted correctly (e.g., does the question-answering dictionary contain the key "answer", is sentiments from textClassify containing the labels "positive" and "negative").

(boolean; only for "fine-tuned"-models) Whether to return all prediction scores or just the one of the predicted class.

(string; only for "fine-tuned"-models) The function to apply to the model outputs to retrieve the scores.

(Integer; only for "fine-tuned" models) Set seed.

Setting from stats::predict can be called.

Object containing several models.

List of word embeddings (if using word embeddings from more than one text-variable use dim_names = TRUE throughout the pipeline).

A tibble/dataframe with additional variables used in the training of the models (optional).

Settings from textPredict.

The observed scores (i.e., what was used to predict when training a model).

The second observed scores (default = NULL; i.e., for when comparing models that are predicting different outcomes. In this case a bootstrap procedure is used to create two distributions of correlations that are compared (see description above).

The predicted scores from model 1.

The predicted scores from model 2 that will be compared with model 1.

Character string specifying the comparison approach.

- "t-test": Use when comparing prediction errors from two models predicting the same outcome (only supported when statistic = "correlation"). Performs a paired t-test on absolute errors.

- "bootstrap_difference": Use to compare AUCs from two models predicting either the same outcome (e.g., does Model A outperform Model B on the same classification task) or different outcomes (e.g., mental vs physical health). Bootstraps the difference in AUC across resamples to compute confidence intervals and p-values.

- "bootstrap_overlap": Use when comparing predictions for different outcomes by generating bootstrap distributions of correlation or AUC values (depending on statistic), and testing overlap in distributions. (requires the overlapping package).

Choose the method that aligns with your research question: error comparison on the same outcome, AUC difference testing on the same or different outcomes, or overlap of performance distributions across different outcomes.

Character ("correlation", "auc") describing statistic to be compared in bootstrapping.

Paired test or not in stats::t.test (default TRUE).

Character "first" or "second" for computing the auc in the bootstrap.

Number of bootstraps (when providing y2).

Set seed.

Settings from stats::t.test or overlapping::overlap (e.g., plot = TRUE).

(character) Word or text variable to be plotted.

Word embeddings from textEmbed for the words to be plotted (i.e., the aggregated word embeddings for the "words" parameter).

Word embeddings from textEmbed for individual words (i.e., decontextualized embeddings).

Numeric variable that the words should be plotted according to on the x-axes.

Numeric variable that the words should be plotted according to on the y-axes (default = NULL, i.e., a 1-dimensional plot is created).

Number of PCA dimensions applied to the word embeddings in the beginning of the function (default = NULL). A number below 1 takes out % of variance; An integer specify number of components to extract. (default is NULL as this setting has not yet been evaluated).

(character) Method to aggregate the word embeddings (default = "mean"; see also "min", "max", and "[CLS]").

(character) Method to split the axes (default = "quartile" involving selecting lower and upper quartile; see also "mean"). However, if the variable is only containing two different values (i.e., being dichotomous) mean split is used.

Compute the power of the frequency of the words and multiply the word embeddings with this in the computation of aggregated word embeddings for group low (1) and group high (2). This increases the weight of more frequent words.

(numeric) Option to select words that have occurred a specified number of times (default = 0); when creating the Supervised Dimension Projection line (i.e., single words receive Supervised Dimension Projection and p-value).

(boolean) Separately mean centering the Group 1 split aggregation embedding, and the Group 2 split aggregation embedding

(boolean) Separately mean centering the G1 and G2 split aggregation embeddings

(numeric) Number of permutations in the creation of the null distribution (default = 10000).

(numeric) Setting to split Npermutations to avoid reaching computer memory limits; set it lower than Npermutations <- and the higher it is set the faster the computation completes, but too high may lead to abortion (default = 50000).

(numeric) Set different seed (default = 1003).

Dataframe from textProjection

Select the k most frequent words to significance test (k = sqrt(100*N); N = number of participant responses). Default = TRUE.

Select words to significance test that have occurred at least min_freq_words_test (default = 1).

Select words to plot that has occurred at least min_freq_words_plot times.

Select number of significant words in each square of the figure to plot. The significant words, in each square is selected according to most frequent words.

Number of significant words to plot on each(positive and negative) side of the x-axes and y-axes, (where duplicates are removed); selects first according to lowest p-value and then according to frequency. Hence, on a two dimensional plot it is possible that plot_n_words_p = 1 yield 4 words.

Number of words that are extreme on Supervised Dimension Projection per dimension. (i.e., even if not significant; per dimensions, where duplicates are removed).

Number of words based on being most frequent. (i.e., even if not significant).

Number of words plotted that are in the middle in Supervised Dimension Projection score (i.e., even if not significant; per dimensions, where duplicates are removed).

(numeric) select random words to plot.

Color for all the titles (default: "#61605e")

If TRUE, also plotting on the y-axes (default is FALSE). Also plotting on y-axes produces a two dimension 2-dimensional plot, but the textProjection function has to have had a variable on the y-axes.

Alpha (default = .05).

(boolean) Allow overlapping (TRUE) or disallow (FALSE) (default = TRUE).

Method to adjust/correct p-values for multiple comparisons (default = "holm"; see also "none", "hochberg", "hommel", "bonferroni", "BH", "BY", "fdr").

(character) Metric to plot according to; "dot_product" or "cohens_d".

Title (default " ")

Label on the x-axes.

Label on the y-axes.

Manually set the length of the x-axes (default = NULL, which uses ggplot2::scale_x_continuous(limits = scale_x_axes_lim); change e.g., by trying c(-5, 5)).

Manually set the length of the y-axes (default = NULL; which uses ggplot2::scale_y_continuous(limits = scale_y_axes_lim); change e.g., by trying c(-5, 5)).

Font type (default: NULL).

The different colors of the words. Note that, at the moment, two squares should not have the exact same colour-code because the numbers within the squares of the legend will then be aggregated (and show the same, incorrect value). (default: c("#398CF9", "#60A1F7", "#5dc688", "#e07f6a", "#EAEAEA", "#40DD52", "#FF0000", "#EA7467", "#85DB8E")).

Vector with minimum and maximum font size (default: c(3, 8)).

Jitter height (default: .0).

Jitter width (default: .03).

Size of the points indicating the words' position (default: 0.5).

Transparency of the lines between each word and point (default: 0.1).

Size of the points not linked with a words (default is to not show it, i.e., 0).

Transparency of the points not linked with a words (default is to not show it, i.e., 0).

Title on the color legend (default: "(SDP)".

Label on the color legend (default: "(x)".

Label on the color legend (default: "(y)".

Position on the x coordinates of the color legend (default: 0.02).

Position on the y coordinates of the color legend (default: 0.05).

Height of the color legend (default 0.15).

Width of the color legend (default 0.15).

Font size (default: 7).

Font size of the values in the legend (default: 2).

(string) Colour of the numbers in the box legend.

Shows a point representing the aggregated word embedding for group 1 (default = FALSE).

Shows a point representing the aggregated word embedding for group 2 (default = FALSE).

Shows a point representing the aggregated direction embedding (default = FALSE).

Size of the points representing the group_embeddings1, group_embeddings2 and projection_embedding

Shape type of the points representing the group_embeddings1, group_embeddings2 and projection_embeddingd

Color

Color

Color

Set different seed.

Explore where specific words are positioned in the embedding space. For example, c("happy content", "sad down").

Specify the color(s) of the words being explored. For example c("#ad42f5", "green")

Specify the names of the point for the aggregated word embeddings of all the explored words.

Specify how to aggregate the word embeddings of the explored words.

manually remove words from the plot (which is done just before the words are plotted so that the remove_words are part of previous counts/analyses).

Set color to words that have higher frequency (N) on the other opposite side of its dot product projection (default = NULL).

Remove words that have higher frequency (N) on the other opposite side of its dot product projection (default = FALSE).

Provide a semantic space if using static embeddings and wanting to explore words.

Scaling word embeddings before aggregation.

(string) A question

(string) The context(s) where the model will look for the answer.

(string) HuggingFace name of a pre-trained language model that have been fine-tuned on a question answering task.

(string) Device to use: 'cpu', 'gpu', or 'gpu:k' where k is a specific device number

(boolean) If TRUE this will turn on tokenizer parallelism.

(string) Set the logging level. Options (ordered from less logging to more logging): critical, error, warning, info, debug

(boolean) Stop returning some incorrectly formatted/structured results. This setting does CANOT evaluate the actual results (whether or not they make sense, exist, etc.). All it does is to ensure the returned results are formatted correctly (e.g., does the question-answering dictionary contain the key "answer", is sentiments from textClassify containing the labels "positive" and "negative").

(integer) (int) Indicates number of possible answer span(s) to get from the model output.

(integer) If the context is too long to fit with the question for the model, it will be split into overlapping chunks. This setting controls the overlap size.

(integer) Max answer size to be extracted from the model’s output.

(integer) The max total sentence length (context + question) in tokens of each chunk passed to the model. If needed, the context is split in chunks (using doc_stride as overlap).

(integer) The max question length after tokenization. It will be truncated if needed.

(boolean) Whether or not impossible is accepted as an answer.

(Integer) Set seed.

Word embeddings from textEmbed().

Word embeddings from textEmbed().

(character) Character string describing type of measure to be computed. Default is "cosine" (see also "spearmen", "pearson" as well as measures from textDistance() (which here is computed as 1 - textDistance) including "euclidean", "maximum", "manhattan", "canberra", "binary" and "minkowski").

(boolean; from base::scale) If center is TRUE then centering is done by subtracting the column means (omitting NAs) of x from their corresponding columns, and if center is FALSE, no centering is done.

(boolean; from base::scale) If scale is TRUE then scaling is done by dividing the (centered) columns of x by their standard deviations if center is TRUE, and the root mean square otherwise.

Word embeddings from textEmbed().

(character) Character string describing type of measure to be computed. Default is "cosine" (see also "spearmen", "pearson" as well as measures from textDistance() (which here is computed as 1 - textDistance) including "euclidean", "maximum", "manhattan", "canberra", "binary" and "minkowski").

(boolean; from base::scale) If center is TRUE then centering is done by subtracting the column means (omitting NAs) of x from their corresponding columns, and if center is FALSE, no centering is done.

(boolean; from base::scale) If scale is TRUE then scaling is done by dividing the (centered) columns of x by their standard deviations if center is TRUE, and the root mean square otherwise.

Word embeddings from textEmbed().

Word embedding from textEmbed (from only one text).

(character) Character string describing type of measure to be computed. Default is "cosine" (see also "spearmen", "pearson" as well as measures from textDistance() (which here is computed as 1 - textDistance) including "euclidean", "maximum", "manhattan", "canberra", "binary" and "minkowski").

(boolean; from base::scale) If center is TRUE then centering is done by subtracting the column means (omitting NAs) of x from their corresponding columns, and if center is FALSE, no centering is done.

(boolean; from base::scale) If scale is TRUE then scaling is done by dividing the (centered) columns of x by their standard deviations if center is TRUE, and the root mean square otherwise.

(string) A variable or a tibble/dataframe with at least one character variable.

(explicit integer; e.g., 10L) The minimum number of tokens in the summed output.

(explicit integer higher than min_length; e.g., 20L) The maximum number of tokens in the summed output.

(string) Specififcation of a pre-trained language model that have been fine-tuned on a summarization task, such as ’bart-large-cnn’, ’t5-small’, ’t5-base’, ’t5-large’, ’t5-3b’, ’t5-11b’.

(string) Device to use: 'cpu', 'gpu', or 'gpu:k' where k is a specific device number.

(boolean) If TRUE this will turn on tokenizer parallelism.

(string) Set the logging level. Options (ordered from less logging to more logging): critical, error, warning, info, debug

(boolean) Stop returning some incorrectly formatted/structured results. This setting does CANOT evaluate the actual results (whether or not they make sense, exist, etc.). All it does is to ensure the returned results are formatted correctly (e.g., does the question-answering dictionary contain the key "answer", is sentiments from textClassify containing the labels "positive" and "negative").

(boolean) Whether or not the outputs should include the decoded text.

(boolean) Whether or not the output should include the prediction tensors (as token indices).

(boolean) Option to clean up the potential extra spaces in the returned text.

(Integer) Set seed.

A character variable or a tibble/dataframe with at least one character variable.

Character string specifying pre-trained language model (default 'bert-base-uncased'). For full list of options see pretrained models at HuggingFace. For example use "bert-base-multilingual-cased", "openai-gpt", "gpt2", "ctrl", "transfo-xl-wt103", "xlnet-base-cased", "xlm-mlm-enfr-1024", "distilbert-base-cased", "roberta-base", or "xlm-roberta-base".

(numeric) Maximum number of tokens in a string to handle before switching to embedding text sentence by sentence.

Name of device to use: 'cpu', 'gpu', 'gpu:k' or 'mps'/'mps:k' for MacOS, where k is a specific device number.

If TRUE this will turn on tokenizer parallelism. Default FALSE.

The maximum length (in number of tokens) for the inputs to the transformer model (default the value stored for the associated model).

Set to TRUE if the accessed model is gated.

The token needed to access the gated model. Create a token from the ['Settings' page](https://huggingface.co/settings/tokens) of the Hugging Face website. An an environment variable HUGGINGFACE_TOKEN can be set to avoid the need to enter the token each time.

use a model with custom code on the Huggingface Hub

Set the logging level. Default: "warning". Options (ordered from less logging to more logging): critical, error, warning, info, debug

(string) Language to tokenise and count.

(numeric) Threshold deciding which words to remove

(tibble/data.frame) A tibble with a text-variable to be analysed, and optional numeric/categorical variables that you might want to use for later analyses testing the significance of topics in relation to these variables.

(string) Name of the text-variable in the data tibble that you want to perform topic modeling on.

(string) Name of the embedding model to use such as "miniLM", "mpnet", "multi-mpnet", "distilroberta".

(string) The dimension reduction algorithm, currently only "default" is supported.

(string) The clustering algorithm to use, currently only "default" is supported.

(string) Name of the vectorizer model, currently only "default" is supported.

(string) Name of the representation model used for topics, including "keybert" or "mmr".

(integer) Determine the number of top words presented for each topic.

(vector) Two-dimensional vector indicating the ngram range used for the vectorizer model.

(string) Name of the stopword dictionary to use.

(integer) The minimum document frequency of terms.

(boolean) Determine whether bm25_weighting is used for ClassTfidfTransformer.

(boolean) Determine whether frequent words are reduced by ClassTfidfTransformer.

(integer) The random seed for initialization of the umap model.

(string) The directory for saving results.

(tibble/data.frame) A tibble with a text-variable to be analysed, and optional numeric/categorical variables that you might want to use for later analyses testing the significance of topics in relation to these variables.

(string) Name of the text-variable in the data tibble that you want to perform topic modeling on.

(string) The dimension reduction algorithm, currently only "default" is supported.

(string) The clustering algorithm to use, currently only "default" is supported.

(string) The directory for saving results.

(string) Name of the embedding model to use such as "miniLM", "mpnet", "multi-mpnet", "distilroberta".

(list) The trained model

(string) The x variable name to be predicted, and to be plotted (only needed for regression or correlation)

(string) The y variable name to be predicted, and to be plotted (only needed for regression or correlation)

(vector) The control variables (not supported yet)

(string) The test method to use, either "correlation","t-test", "linear_regression","logistic_regression", or "ridge_regression"

(character) Method to adjust/correct p-values for multiple comparisons (default = "none"; see also "holm", "hochberg", "hommel", "bonferroni", "BH", "BY", "fdr").

Parameter settings from topicsTest in the topics-package.

(list) The output from textTopics.

(tibble/data.frame) A tibble with the data

(string) The name of the text variable that the topic model was trained on

(list) A trained topics model. For examples from topicsModel(). Should be NULL if plotting ngrams.

(list) The output from the the topicsGram() function . Should be NULL if plotting topics.

(list) The test results; if plotting according to dimension(s) include the object from topicsTest() function.

(integer) The seed to set for reproducibility; need to be the same seed number as in in

Parameters from the topicsPlot() function in the topics package.

Word embeddings from textEmbed (or textEmbedLayerAggreation). Can analyze several variables at the same time; but if training to several outcomes at the same time use a tibble within the list as input rather than just a tibble input (i.e., keep the name of the wordembedding).

Numeric variable to predict. Can be several; although then make sure to have them within a tibble (this is required even if it is only one outcome but several word embeddings variables).

Default is "automatic", so if y is a factor random_forest is used, and if y is numeric ridge regression is used. This can be overridden using "regression" or "random_forest".

Arguments from textTrainRegression or textTrainRandomForest the textTrain function.

Word embeddings from textEmbed (or textEmbedLayerAggreation). It is possible to have word embeddings from one text variable and several numeric/categorical variables; or vice verse, word embeddings from several text variables to one numeric/categorical variable. It is not possible to mix numeric and categorical variables.

Tibble with several numeric or categorical variables to predict. Please note that you cannot mix numeric and categorical variables.

(character) Default is "automatic"; see also "regression" and "random_forest".

(character) Option not to save all output; default "all". See also "only_results" and "only_results_predictions".

(character) A character string describing type of correlation (default "Pearson").

(character) Type of evaluative measure to assess models on (default "rmse").

Method to adjust/correct p-values for multiple comparisons. (default = "holm"; see also "none", "hochberg", "hommel", "bonferroni", "BH", "BY", "fdr").

Arguments from textTrainRegression or textTrainRandomForest (the textTrain function).

Word embeddings from textEmbed (or textEmbedLayerAggregation). If several word embedding are provided in a list they will be concatenated.

Numeric variable to predict.

(numeric) Numeric vector that specifies the percentages of the total number of data points to include in each sample (default = c(25,50,75,100), i.e., correlations are evaluated for 25 each new sample.

Determine whether to use a list of word embeddings or an individual word_embedding (default = "individually", also "concatenate"). If a list of word embeddings are provided, then they will be concatenated.

(numeric) Value that determines the number of times to repeat the cross-validation (i.e., number of tests). (default = 1, i.e., cross-validation is only performed once). Warning: The training process gets proportionately slower to the number of cross-validations, resulting in a time complexity that increases with a factor of n (n cross-validations).

Sample a "random" sample for each subset from all data or sample a "subset" from the larger subsets (i.e., each subset contain the same data).

If TRUE it only searches the penalty and mixture search grid once, and then use the same thereafter; if FALSE, it searches the grid every time.

Type of model. Default is "regression"; see also "logistic" and "multinomial" for classification.

(numeric) Hyper parameter that is tuned (default = 10^seq(-16,16)).

A number between 0 and 1 (inclusive) that reflects the proportion of L1 regularization (i.e. lasso) in the model (for more information see the linear_reg-function in the parsnip-package). When mixture = 1, it is a pure lasso model while mixture = 0 indicates that ridge regression is being used (specific engines only).

(numeric) Set different seed (default = 2024).

Additional parameters from textTrainRegression.

(list) One or several objects returned by the function textTrainN as a list (e.g, list(object1, object2)). Also, if several models are provided, then one can add a vector c() with settings (i.e the parameters below) for each model (make sure to add the settings in the order as the models are ordered, if you look to keep the original settings then write "").

(numeric) Vector containing the percents of the total number of data points that is included in each sample (default = NULL, which takes the breaks from the percentages). If several models are provided, then one can add a vector c() with settings for each model (make sure to add the settings in the order as the models are ordered).

(character, "percent" or "quantity") Determines whether the x-axis-values should represent the number of elements in each sample, or the number of percent of the total data they represent (default = "percent").

(numeric) Optional. Determines the y_range. E.g, y_range = c(1,2) sets the y_range from 1 to 2 (default = NULL).

(character) Determine plot title (default = "Cross-validated correlation coefficients across different sample sizes").

(character) Determine x-axis-label (default = "Sample Size (percent)").

(character) Determine y-axis-label (default = "Correlation Coefficient (r)").

(character, (Hex color codes)) Determine point color (default = "#5dc688"). Can set a vector if several results_data are provided.

Default "std_err"; see also "std", NULL. Can set a vector if several results_data are provided.

(character, (Hex color codes)) Determine error-bar color (default = "#60A1F7"). Can set a vector if several results_data are provided.

(character, (Hex color codes)) Determine line color (default = "grey"). Can set a vector if several results_data are provided.

(numeric) Determine bar-width (default = 1). Can set a vector if several results_data are provided.

(numeric) Determine bar-size (default = 1). Can set a vector if several results_data are provided.

(numeric) Determine line-size (default = 1). Can set a vector if several results_data are provided.

(character, either "straight" or "smooth") Determine line-type (default = "straight"). Can set a vector if several results_data are provided.

(numeric) Determine points size (default = 1). Can set a vector if several results_data are provided.

(boolean) Determine wether to log-transform x in case of displaying number of samples (default = FALSE).

Word embeddings from textEmbed.

Categorical variable to predict.

(optional) Variables to be appended after the word embeddings (x); if wanting to preappend them before the word embeddings use the option first = TRUE. If not wanting to train with word embeddings, set x_append = NULL (default = null).

(boolean) Option to add variables before or after all word embeddings (default = FALSE).

(character) Cross-validation method to use within a pipeline of nested outer and inner loops of folds (see nested_cv in rsample). Default is using cv_folds in the outside folds and "validation_split" using rsample::validation_split in the inner loop to achieve a development and assessment set (note that for validation_split the inside_folds should be a proportion, e.g., inside_folds = 3/4); whereas "cv_folds" uses rsample::vfold_cv to achieve n-folds in both the outer and inner loops.

(numeric) Number of folds for the outer folds (default = 10).

(numeric) Number of folds for the inner folds (default = 3/4).

(string or tibble; default "y") Variable to stratify according; if a string the variable needs to be in the training set - if you want to stratify according to another variable you can include it as a tibble (please note you can only add 1 variable to stratify according). Can set it to NULL.

(boolean) Whether to stratify the outside folds.

(numeric) The number of bins wanted to stratify a numeric stratification variable in the outer cross-validation loop (default = 4).

(boolean) Whether to stratify the outside folds.

The number of bins wanted to stratify a numeric stratification variable in the inner cross-validation loop (default = 4).

Default is "classification" ("regression" is not supported yet).

(boolean) Normalizes dimensions to have a mean of zero; default is set to FALSE For more info see (step_center in recipes).

(boolean) Normalizes dimensions to have a standard deviation of one; default is set to FALSE. For more info see (step_scale in recipes).

Pre-processing threshold for PCA. Can select amount of variance to retain (e.g., .90 or as a grid c(0.80, 0.90)); or number of components to select (e.g., 10). (To skip this step, set preprocess_PCA to NA) Default is "min_halving", which is a function that selects the number of PCA components based on number of participants and feature (word embedding dimensions) in the data. The formula is: preprocess_PCA = round(max(min(number_features/2), number_participants/2), min(50, number_features))).

Default is "extratrees", which thus implement a random forest; can also select: NULL, "gini" or "hellinger"; if these are selected your mtry settings will be overridden (see Geurts et al. (2006) Extremely randomized trees for details; and see the ranger r-package for details on implementations).

Hyper parameter that may be tuned; default: c(1, 20, 40),

Hyper parameter that may be tuned; default: c(1, 20, 40)

Number of trees to use (default 1000).

If several results are tied for different parameters (i.e., mtry or min_n), then select the "lowest_mtry", "highest_mtry", "median_mtry", or "lowest_min_n", the "highest_min_n" or the "median_min_n" order of all the tied mtry/min_n

(character) Measure to evaluate the models in order to select the best hyperparameters default "roc_auc"; see also "accuracy", "bal_accuracy", "sens", "spec", "precision", "kappa", "f_measure".

(character) Text to describe your model (optional; good when sharing the model with others).

If TRUE it enables the use of multiple cores if the computer system allows for it (i.e., only on unix, not windows). Hence it makes the analyses considerably faster to run. Default is "multi_cores_sys_default", where it automatically uses TRUE for Mac and Linux and FALSE for Windows. Note that having it to TRUE does not enable reproducable results at the moment (i.e., cannot set seed).

(character) Option not to save all output; default "all". See also "only_results" and "only_results_predictions".

(Boolean) From fisher.test: a logical indicating whether to compute p-values by Monte Carlo simulation, in larger than 2 × 2 tables. The test can be turned off if set to "turn_off".

(numeric) Set different seed (default = 2020).

For example settings in yardstick::accuracy to set event_level (e.g., event_level = "second").

Word embeddings from textEmbed (or textEmbedLayerAggregation). If several word embedding are provided in a list they will be concatenated.

Numeric variable to predict.

(optional) Variables to be appended after the word embeddings (x); if wanting to preappend them before the word embeddings use the option first = TRUE. If not wanting to train with word embeddings, set x = NULL (default = NULL).

(boolean) Option to add variables before or after all word embeddings (default = False).

(character) Cross-validation method to use within a pipeline of nested outer and inner loops of folds (see nested_cv in rsample). Default is using "cv_folds" in the outside folds and "validation_split" using rsample::validation_split in the inner loop to achieve a development and assessment set (note that for "validation_split" the inside_folds should be a proportion, e.g., inside_folds = 3/4); whereas "cv_folds" uses rsample::vfold_cv to achieve n-folds in both the outer and inner loops. Use "group_cv" to ensure that all cases with the same ID remain in the same fold. (it uses rsample::group_vfold_cv uses to ensure that all cases with the same ID remain in the same fold. group_vfold_cv cannot handle stratification, so if that is requested, it tries to approximate stratification while preserving group integrity.

(variable) If specifying cv_method = "group_cv", you need to submit an id variable here.

(numeric) Number of folds for the outer folds (default = 10).

(numeric) The proportion of data to be used for modeling/analysis; (default proportion = 3/4). For more information see validation_split in rsample.

(string or tibble; default "y") Variable to stratify according; if a string the variable needs to be in the training set - if you want to stratify according to another variable you can include it as a tibble (please note you can only add 1 variable to stratify according). Can set it to NULL.

(boolean) Whether to stratify the outside folds.

(numeric) The number of bins wanted to stratify a numeric stratification variable in the outer cross-validation loop (default = 4).

Whether to stratify the outside folds.

The number of bins wanted to stratify a numeric stratification variable in the inner cross-validation loop (default = 4).

Type of model. Default is "regression"; see also "logistic" and "multinomial" for classification.

(character) Type of evaluative measure to select models from. Default = "rmse" for regression, "weighted_correlations for weighted regression and "bal_accuracy" for logistic. For regression use "rsq" or "rmse"; and for classification use "accuracy", "bal_accuracy", "sens", "spec", "precision", "kappa", "f_measure", or "roc_auc", (for more details see the yardstick package). See also the method_cor setting below.

(boolean) If TRUE, the aggregated word embeddings (mean, min, and max) are saved for comparison with other language input when the model is applied to other types of data.

(Character column) If you provide the raw language data used for making the embeddings, the language distribution (i.e., a word and frequency table) will be saved to the model object. This enables calculating similarity scores when the model is being applied to new language domains. Note that this saves the individual words, which, if you are analyzing sensitive data, can be problematic from a privacy perspective; to some extent this can be mitigated by increasing the number of words needed to be saved.

(numeric) Minimum number a words need to occur in the data set to be saved to the language distribution.

(boolean) Normalizes dimensions to have a mean of zero; default is set to TRUE. For more info see (step_center in recipes).

(boolean) Normalize dimensions to have a standard deviation of one; default is set to TRUE. For more info see (step_scale in recipes).

Pre-processing threshold for PCA (to skip this step set it to NA). Can select amount of variance to retain (e.g., .90 or as a grid c(0.80, 0.90)); or number of components to select (e.g., 10). Default is "min_halving", which is a function that selects the number of PCA components based on number of participants and feature (word embedding dimensions) in the data. The formula is: preprocess_PCA = round(max(min(number_features/2), number_participants/2), min(50, number_features))).

(numeric) Hyper parameter that is tuned (default = 10^seq(-16,16)).

If several results are tied for different parameters (i.e., penalty or mixture), then select the "lowest_penalty", "highest_penalty", "median_penalty", or "lowest_mixture", the "highest_mixture" or the "median_mixture" order of all the tied penalties/mixtures.

A number between 0 and 1 (inclusive) that reflects the proportion of L1 regularization (i.e. lasso) in the model (for more information see the linear_reg-function in the parsnip-package). When mixture = 1, it is a pure lasso model while mixture = 0 indicates that ridge regression is being used (specific engines only).

By default this setting is turned off (i.e., NA). To use this method, set it to the highest number of predictors you want to test. Then the X first dimensions are used in training, using a sequence from Kjell et al., 2019 paper in Psychological Methods. Adding 1, then multiplying by 1.3 and finally rounding to the nearest integer (e.g., 1, 3, 5, 8). This option is currently only possible for one embedding at the time.

Default FALSE (can be set to TRUE if something else than word_embeddings are trained).

Type of correlation used in final model estimation evaluation (default "pearson"; can set to "spearman" or "kendall").

(character) Text to describe your model (optional; good when sharing the model with others).

If TRUE it enables the use of multiple cores if the computer system allows for it (i.e., only on unix, not windows). Hence it makes the analyses considerably faster to run. Default is "multi_cores_sys_default", where it automatically uses TRUE for Mac and Linux and FALSE for Windows.

(character) Option not to save all output; default = "all". see also "no_plot", "only_results", and "only_results_predictions". Note that "no_plot" is good when wanting to save a logistic or multnomial regression, since the lot makes the saved object bloated when being saved.

(Boolean or string) From fisher.test: a logical indicating whether to compute p-values by Monte Carlo simulation, in larger than 2 * 2 tables. The test can be turned off if set to "turn_off".

(numeric) Set different seed (default = 2020).

Optional vector containing weights (default = NULL); for details see importance_weights hardhat. For now only working for model = "regression".

For example settings in yardstick::accuracy to set event_level (e.g., event_level = "second").

(string) The text to be translated.

(string) The input language. Might be needed for multilingual models (it will not have any effect for single pair translation models). using ISO 639-1 Code, such as: "en", "zh", "es", "fr", "de", "it", "sv", "da", "nn".

(string) The desired language output. Might be required for multilingual models (will not have any effect for single pair translation models).

(string) Specify a pre-trained language model that have been fine-tuned on a translation task.

(string) Name of device to use: 'cpu', 'gpu', or 'gpu:k' where k is a specific device number

(boolean) If TRUE this will turn on tokenizer parallelism.

(string) Set the logging level. Options (ordered from less logging to more logging): critical, error, warning, info, debug

(boolean) Stop returning some incorrectly formatted/structured results. This setting does CANOT evaluate the actual results (whether or not they make sense, exist, etc.). All it does is to ensure the returned results are formatted correctly (e.g., does the question-answering dictionary contain the key "answer", is sentiments from textClassify containing the labels "positive" and "negative").

(boolean) Whether or not to include the predictions' tensors as token indices in the outputs.

(boolean) Whether or not to also output the decoded texts.

(boolean) Whether or not to clean the output from potential extra spaces.

(Integer) Set seed.

Set max length of text to be translated

(string) The sequence(s) to classify (not that they will be truncated if the model input is too large).

(string) The set of class labels that is possible in the to classification of each sequence. It may be a single label, a string of comma-separated labels, or a list of labels.

(string; optional) The template that is used for turning each of the label into an NLI-style hypothesis. This template must include a "" or similar syntax so that the candidate label can be inserted into the template. For example, the default template is "This example is ." With the candidate label "sports", this would be fed into the model like "<cls> sequence to classify <sep> This example is sports . <sep>". The default template works well in many cases, but it may be worthwhile to experiment with different templates depending on the task setting (see https://huggingface.co/docs/transformers/).

(boolean; optional) It indicates whether multiple candidate labels can be true. If FALSE, the scores are normalized such that the sum of the label likelihoods for each sequence is 1. If TRUE, the labels are considered independent and probabilities are normalized for each candidate by doing a softmax of the entailment score vs. the contradiction score.

(string) Specify a pre-trained language model that have been fine-tuned on a translation task.

(string) Name of device to use: 'cpu', 'gpu', or 'gpu:k' where k is a specific device number

(boolean) If TRUE this will turn on tokenizer parallelism.

(string) Set the logging level. Options (ordered from less logging to more logging): critical, error, warning, info, debug

(boolean) Stop returning some incorrectly formatted/structured results. This setting does CANOT evaluate the actual results (whether or not they make sense, exist, etc.). All it does is to ensure the returned results are formatted correctly (e.g., does the question-answering dictionary contain the key "answer", is sentiments from textClassify containing the labels "positive" and "negative").

(Integer) Set seed.

the full path to the Python executable, for which text required python packages is installed.

set a path to the Python virtual environment with text required python packages installed Example: virtualenv = "~/myenv"

set a path to the anaconda virtual environment with text required python packages installed Example: condalenv = "myenv"

logical; if FALSE, use the first text required python packages installation found; if TRUE, list available text required python packages installations and prompt the user for which to use. If another (e.g. python_executable) is set, then this value will always be treated as FALSE.

logical; if TRUE, text will ignore the saved settings in the profile and initiate a search of new settings.

logical; if TRUE, the current text required python packages setting will be saved for the future use.

logical; check whether conda/virtual environment generated by textrpp_install() exists

logical; Test whether function (textEmbed) that requires python packages works.

logical; asking whether user wants to set the environment as default.

character; path to conda executable. Default "auto" which automatically find the path

Boolean; update to the latest version of Miniconda after install? (should be combined with force_conda = TRUE)

Boolean; force re-installation if Miniconda is already installed at the requested path?

character; default is "rpp_version_system_specific_defaults", because diffent systems require different combinations of python version and packages. It is also possible to specify your own, such as c("torch==2.0.0", "transformers==4.19.2", "numpy", "pandas", "nltk", "scikit-learn", "datasets", "evaluate").

character; default is "python_version_system_specific_defaults". You can specify your Python version for the condaenv yourself. installation.

character; name of the conda-environment to install text required python packages. Default is "textrpp_condaenv".

TRUE to use pip for installing rpp If FALSE, conda package manager with conda-forge channel will be used for installing rpp.

character; path to Python only for virtualenvironment installation

logical; ask whether to proceed during the installation

character;

character; e.g., "python", only for virtualenvironment installation

path to conda executable, default to "auto" which automatically finds the path

logical; ask whether to proceed during the installation

character; name of conda environment to remove