estimateSamplingUnit: estimateSamplingUnit

In j2scode/predictifyR: Word Prediction Language Model Evaluation

Description

estimateSamplingUnit Estimates the sampling unit for corpus sampling

Usage

estimateSamplingUnit(korpus, sampleSizes = c(100, 500, 1000, 2000),
  numSamples = 30)

Arguments

korpus	List containing the meta data for the corpus
sampleSizes	Integer vector of sample sizes to be evaluated
numSamples	Integer indicating number of samples to evaluate

Details

This function takes as its parameters, the korpus meta data and the POS tags selected for this analysis and compares the distributions of lexical features across pairs of samples of varying sizes. The results of chi-squared tests for selected features are averaged over the samples. The function returns a data frame indicating average chi-squared p-values for each feature and sampling unit size.

Value

analysis A list containing:

• sampleSizeSample size being tested

• scores(long)Long dataframe of chi-squared scores at various sample sizes

• scores(wide)Wide dataframe of chi-squared scores at various sample sizes

Author(s)

John James, [e-mail]

See Also

analyzeLexicalFeatures text2spc.fnc lnre lnre.spc N V EV chisq.test

Other sample size estimate functions: estimateCorpusSize, estimateRegisterSize, estimateSampleSize

j2scode/predictifyR documentation built on May 14, 2019, 10:34 a.m.