Implementation of PCA & ICA for voice ecognition and separation of speech

Principle Component Analysis is great to evaluate the correlation among variable and reduce data dimensionally without loss of any data. The ability of analyzing the property of voice, reducing noises and extracting the valuable data of voice makes PCA an integral part of voice recognition. In digital signal processing signal estimation is required; signal may be superimposed by several interfering sources. To find one desired source signal Independent Component Analysis can be implemented. ICA recovers a set of independent signal from a set of measured signals by using statistical analysis of signal.