In polvanrijn/contouR: Compute dynamic features on pitch contours

First set some variables

pt_path = "/Users/pol/owncloud/HiWi_folder/02_running_projects/02_MA_thesis/01_Data/Preprocessed/PitchTiers/08_manual_corrected/"
pt_file_names = list.files(pt_path)
pt_file_names = pt_file_names[grepl(".PitchTier", pt_file_names)]

library(contouR)

Now get the distributions, we'll compute skewness, kurtosis, entropy and standard deviation

results = compute_distribution_features(pt_file_names, pt_path)
head(results) # Let's have a look at the first 6

We can visualize some correlations

library(corrplot)
corrplot(cor(results[, 2:5]), method="circle", type = "upper")

Show some boxplots

results$filename = results$name

interesting_columns = names(results)[2:5]
for (col in interesting_columns){
  print(significance_test(na.omit(results), col))
}

Now we need to save the features

write.csv(results, "distribution_features.csv", row.names = FALSE)

polvanrijn/contouR documentation built on Feb. 10, 2020, 3:45 a.m.