Study of machine learning algorithms for Sinhala speech recognition

Abstract

Speech is the primary mode of communication among humans and the most natural and efficient form of exchanging information. Therefore, it is logical that the next technological development in natural language speech recognition for Human Computer Interaction is, Artificial Intelligence. Speech recognition can be defined as the process of converting speech signal to a sequence of words by an algorithm implemented using a computer program. Speech processing is one of the challenging areas of signal processing. The main objective of the study was to conduct a study on speech recognition approaches to improve the accuracy level of Sinhala speech recognition. This study was conducted in order to find the optimal algorithm for accurate Sinhala speech recognition. According to the implementation architecture of speech recognition, feature extraction and the pattern recognition phases can be varied with different algorithms. The study identified that Linear Predictive Coding (LPC) and Hidden Markov Model (HMM) gives most accurate results than other combine algorithms.