Decision Tree Algorithm for Determining Gender based on Sound Recording

Abstrak

Humans are born with their own uniqueness and character, even though having a uniqueness that classifies humans by sex is something that is not difficult to do. In humans, the way to differentiate between men and women is to look at physical differences and listen to different voices between men and women. On the computer, gender differences can also be identified by classifying male and female voices using a tree decision algorithm with a previously appeared dataset in the form of a sample of 3168 male and female voice recordings, the voice recording sample is processed by acoustic analysis in R using Seewave and tuneR packages with frequency interval 0 hz - 280 hz. Meanfun is used as a predictor for the root sound dataset, with a threshold <= 0.142, with an optimal depth value of 6 using the cross validation method, the results achieved are the accuracy training set of 99.18809% and the accuracy test set reaches 95.89905%.