Noise Cancelling for Robust Speaker Identification Using Least Mean Square

Abstract

This study focused on noise cancelling using Least Muns Square (Ll\IS) for robust speaker identifiration. Noise ranrrlling ran be used to oHrcome sptaker idrntification diffirulties in recognizing noisy voicrs. This study used the LMS on data preprocessing. Based on experimental results, it can be concluded that speaker identification using LMS at data preprocessing produces higher arcuracy than without using LMS. The highest accurary of data that using the LMS is 84.64•/e whereas the data without using the LMS only 2.11%.