transition_predictions: Transition Predictions

In AnimalSequences: Analyse Animal Sequential Behaviour and Communication

Transition Predictions

Description

This function takes sequences of elements and uses a machine learning classifier to predict the next elements in the sequence. It supports n-gram tokenization and k-fold cross-validation. Optionally, it can upsample the training data.

Usage

transition_predictions(
  sequences,
  classifier = "nb",
  ngram = 2,
  upsample = TRUE,
  k = 10
)

Arguments

sequences	A list of character strings representing sequences of elements.
classifier	A character string specifying the classifier to use. Options are 'nb' for Naive Bayes and 'forest' for random forest.
ngram	An integer specifying the number of elements to consider in the n-gram tokenization. Default is 2.
upsample	A logical value indicating whether to upsample the training data to balance class distribution. Default is TRUE.
k	An integer specifying the number of folds for k-fold cross-validation. Default is 10.

Value

A list containing the mean accuracy, mean null accuracy, and a data frame of prediction errors.

Examples

sequences <- list("a b c", "b c d", "c d e")
result <- transition_predictions(sequences, classifier = 'nb', ngram = 2, upsample = TRUE, k = 5)
print(result)

AnimalSequences documentation built on Sept. 30, 2024, 9:18 a.m.