quail.analyze

quail.analyze(egg, subjgroup=None, listgroup=None, subjname='Subject', listname='List', analysis=None, position=0, permute=False, n_perms=1000, parallel=False, match='exact', distance='euclidean', features=None, ts=None, n_jobs=-1)

General analysis function that groups data by subject/list number and performs analysis.

Parameters:

eggEgg data object

The data to be analyzed

subjgrouplist of strings or ints

String/int variables indicating how to group over subjects. Must be the length of the number of subjects

subjnamestring

Name of the subject grouping variable

listgrouplist of strings or ints

String/int variables indicating how to group over list. Must be the length of the number of lists

listnamestring

Name of the list grouping variable

analysisstring

This is the analysis you want to run. Can be accuracy, spc, pfr, temporal or fingerprint

positionint

Optional argument for pnr analysis. Defines encoding position of item to run pnr. Default is 0, and it is zero indexed

permutebool

Optional argument for fingerprint/temporal cluster analyses. Determines whether to correct clustering scores by shuffling recall order for each list to create a distribution of clustering scores (for each feature). The “corrected” clustering score is the proportion of clustering scores in that random distribution that were lower than the clustering score for the observed recall sequence. Default is False.

n_permsint

Optional argument for fingerprint/temporal cluster analyses. Number of permutations to run for “corrected” clustering scores. Default is 1000 ( per recall list).

parallelbool

Option to use multiprocessing (this can help speed up the permutations tests in the clustering calculations)

matchstr (exact, best or smooth)

Matching approach to compute recall matrix. If exact, the presented and recalled items must be identical (default). If best, the recalled item that is most similar to the presented items will be selected. If smooth, a weighted average of all presented items will be used, where the weights are derived from the similarity between the recalled item and each presented item.

distancestr

The distance function used to compare presented and recalled items. Applies only to ‘best’ and ‘smooth’ matching approaches. Can be any distance function supported by numpy.spatial.distance.cdist.

n_jobsint

Number of parallel jobs for fingerprint analysis. Default is -1 (all cores). Only used for fingerprint/temporal analyses when joblib is available.

Returns:

resultquail.FriedEgg

Class instance containing the analysis results