Realization of a sound event detection system using controlled learning methods

Today, very different analysis studies related to sound are carried out. These include work areas such as speech recognition, converting speech into text, voice transcription (parsing of sound). When examining the literature related to the studies, an important part of the studies is the studies on learning and classification of sound data in a data set is observed. As is known, there are billions of voices in nature, and studies with classification methods may be insufficient due to the fact that learning is realized through a limited set of data. The main objective of the thesis study is to "reveal a supervised learning model to make inferences of person voices and background sounds on voice data" is to perform. The thesis carried out the work necessary for processing with sound data and machine learning algorithms. The thesis included researches about acoustics and examination methods and supervised machine learning methods.