Effect of Gender Variability in Speaker Recognition

Nowadays, because of the speech recognition system, there is much interaction between humans and machines. The gender identification system is utilised in a wide range of applications, including call centres, robots, artificial intelligence, and security systems. This research describes a novel technique for identifying gender as male or female by extracting information from audio speech. In this study, gender is predicted from an audio data set using neural networks and a logistic regression structure. This methodology gives an accuracy of 97.8% in gender prediction. The most effective ten data features in the proposed approach were first determined. Afterwards, a neural network was developed as a classifier. The performance comparison of the classification and accuracy value was calculated. In addition, the suggested method's accuracy was compared to the classifier accuracy values generated by conventional machine learning techniques. With a 97.8% success rate, the study was successful in predicting gender.