update_model_decision_bias: Update model's decision biases based on labelled input

In hlplab/MVBeliefUpdatr: Fitting, Summarizing, and Visualizing of Multivariate Gaussian Ideal Observers and Adaptors

Update model's decision biases based on labelled input

Description

Returns model with updated decision biases. This updating process is order-sensitive. For each input observation x, a posterior expectation for each category is obtained based on the specified noise and lapse treatment. This expectation is used to calculate the surprisal (in bits) of the category label x_category. The surprisal is weighted by the learning rate beta to increase the (log-odds of the) decision bias for that category.

Usage

update_model_decision_bias_by_one_observation(
  model,
  beta,
  x,
  x_category,
  noise_treatment = if (!is.null(first(model$Sigma_noise))) "marginalize" else "no_noise",
  lapse_treatment = "no_lapses",
  update_prior = T,
  verbose = F
)

Arguments

model	A model object with decision biases.
beta	The learning rate with which decision biases change. Might be integrated into model objects in the future.
x	The cues of single observation.
x_category	The category label(s) of one or more observations.
noise_treatment	Determines whether perceptual noise is considered during categorization, and how. Can be "no_noise", "sample", or "marginalize". If "no_noise", no noise will be applied to the input, and no noise will be assumed during categorization. If "marginalize", average noise (i.e., no noise) will be added to the stimulus, and 'Sigma_noise' is added to Sigma when calculating the likelihood. This simulates the expected consequences for perceptual noise on categorization *in the limit*, i.e, if the input was categorized infinitely many times. If "sample", then noise is sampled and applied to the input, and 'Sigma_noise' is added to Sigma when calculating the likelihood. This simulates the consequence of perceptual noise *on a particular observation*. If "sample" or "marginalize" are chosen, 'Sigma_noise' must be a covariance matrix of appropriate dimensions. (default: "no_noise" if Sigma_noise is NULL, "marginalize" otherwise).
lapse_treatment	Determines whether attentional lapses can occur during which no updating occurs. Can be "no_lapses", "sample", or "marginalize". If "no_lapses", no lapses occur (even if the model specifies a non-zero ‘lapse_rate'), and all observations lead to updating. If "sample" or "marginalize", the lapse rate from ’ the model will be used. For "sample", sampling determines for each observation whether it was a lapse or not. If an observation was a lapse no updating occurs. For "marginalize", 1 - lapse_rate is the proportion of observations that are assumed to be lapsing trials (default: "no_lapses")
update_prior	Should the prior probability of each category be updated along with the decision bias?
verbose	Should more informative output be provided?

Value

A model object.

hlplab/MVBeliefUpdatr documentation built on March 29, 2025, 10:42 p.m.