Analysing human voice and classification of voice frequencies according to smoking effect

Abstract

Speech is the ability that allows people to communicate with each other. The main element of the speech is voice and voice come together to form speech. Therefore the deterioration of voice will weaken the communication between people and then people's social and business life will lead to negativity. The human voice is affected by a variety of diseases and harmful substances such as cigarettes. In this study, the analysis of the fundamental frequency of the human voice and the effects of cigarette on those frequencies has been studied. In the study, the voices of the adult male subjects taken and recorded in the appropriate environment and conditions. The recorded voices are analyzed with signal processing methods. A data set is prepared with numerical value of analyzed voices and this data set is classified by machine learning methods. The voices that are taken from new individuals are predicted according to the result of this classification. Similar to findings observed in diseases of the human voice by methods based on the results obtained in this study, voices will be analyzed and made pathological analysis is intended to detect diseases from the human voice at future. In the study, the fundamental frequency of the voice that also untested literature in men 25–35 years, is classified with machine learning and classification algorithms with between 69.5% and 87.5% of success ratio.