Classifying Philippine-Korean Speech Analysis based on various Machine Learning Heuristics.

Abstract

Every individual can be identified in their native language, and use them as communication expressing an opinion, sharing emotions and feelings through voices, speech and body language.Language is basic and fundamental to us.Numerous languages in terms of variants in the world today speak, a diverse manner of speech technique and method are common practices and a natural way of speaking.This paper presents a classification of Filipino and Korean speech assessment based on various machine learning.Data are learned using several algorithms including NaivesBayes, K-Means, Support Vector Machine and Multilayer Perceptron. The data collected from recorded voices tests and compares with those classifiers. There are 100 respondents and collect recorded audio. A dataset is split into half 50 came from Filipino and 50 came from Korean. The assessment and evaluation are measured based on its accuracy and correctness.Based on the results shown in the Multilayer Perceptron (MLP) figures, the highest and best performance was 99.49 percent, followed by the Support Vector Machine (SVM) with 98.6 percent, and the Naïve Bayes with 97 percent. With 94.78 percent, 74% and K-means fall to the lowest position.