Publication

Abstract : There are various methods by that emotions can be expressed, each emotion varies by unique experiences and understandings. So, to understand emotions conveyed, a capable system that is capable of modeling and capturing other language emotions is needed. In an intelligent learning environment, this investigation aims to look into the recognition of emotions in voice and text images of learner expressions. A suggested improved convolutional neural network long short-term memory algorithm is tested using a simulation experiment to ensure that it performs as expected after analyzing the effectiveness of numerous deep learning-related neural network algorithms. we propose a model to predict the emotions in a student’s speech or text message. Taking our model as a multi-class classification model, we compare two models. The features which are extracted are used to train different ML classifiers, and an LSTM classifier is trained using the same features. We extracted features from text data, and features from speech data and concatenated both models to make a multi-modal that classifies both speech and text data.