Abstract

With the very important development of computing and storage resources, human-machine interfaces become more and more closer to the natural human interaction. In this paper, we present a vocal calculator based on support vector machine learning and classification. Learning techniques used until now, in this research domain, such as Artificial Neural Networks (ANN) are characterized by weak classification rate and learning speed. SVM based classification is a very powerful method which yield good results in many domains such as face and hand writing recognition. The application of such learning method in voice recognition applied on a calculator give very interesting recognition rate.

Keywords: speech recognition, machine learning, Support vector machine.