https://link.springer.com/chapter/10.1007/978-981-15-7394-1_36Human–machine interaction is becoming popular day by day; to interact with machine, speech emotion recognition is as important as human to human interaction. In this research, we demonstrate a speech emotion recognition system which takes speech as input and classify emotions that the speech contains. We choose multilayer perceptron (MLP) classifier to do this task. Features that we have extracted from speech are mel-frequency cepstral coefficients (MFCC), chroma and mel-spectrogram frequency. RADVES dataset has been used and we have got 73% accuracy.