Effective speech emotion recognition using deep learning approaches for Algerian dialect

Speech Emotion Recognition is an active research problem in the field of deep learning. Where the goal is to detect different emotions within audio speech using deep artificial neural networks. In this work, we focus on speech emotion recognition in the Algerian dialect. The paper introduces a new large Algerian speech emotion dataset collected from different Algerian TV shows. The audio signals were manually annotated as happy, angry, neutral, or sad. After the data collection, we applied several classification methods such as machine learning-based models, convolutional neural networks (CNNs), Long Short Term Memory (LSTM) networks, and Bidirectional LSTM (BLSTM). After applying the proposed models on our dataset, the best classification accuracy is 93.34% achieved by the LSTM-CNN model.