Speech Based Emotion Classification using MLP

Advancement in technology, leading the real-world problems to be solved. There is an interesting concept, called Speech Based Emotion Classification. This is getting popular because this work can be able to create the path of recognizing Real Human Behaviour if it is connected with multiple aspects.

Here, Speech Based Emotion Classification is done by using Neural Networks - MLP (Multi-Layer Perceptron)

In this work, the RAVDESS dataset (https://smartlaboratory.org/ravdess/) has been used as the input source (Audio Datasets). Using these datasets, Emotions are classified. In this work, the "Librosa" library has been used to extract features. For additional metrics and implementation "sklearn" libraries have been used. Using MLP - achieved 57% - 60% of accuracy depends on iterations.

Source Code: https://github.com/Rajesh-Research/Speech-Based-Emotion-Classification-using-MLP

Here, the confusion matrix of the Emotions - Ture Emotions Vs Predicted Emotions.