Published Paper Details:

The goal of this study is to investigate how authors ability to recognise and convey emotions through words and expressions changes in an online, adaptive learning environment. The effectiveness of a method developed to improve a convolution neural network's performance using Bidirectional Long Short-Term Memory is verified through a simulated exercise (CNN-BiLSTM). It was decided to do this after carefully examining the deep learning neural network methods already available. Based on the experimental results, the CNN-BiLSTM method provided in this article achieves an accuracy of up to 98.75%, which is at least 3.15 percentage points more than the accuracy of other algorithms. The book "Deep Learning-Based Artificial Emotional Intelligence" delves at the ways in which AI has been used to emotional intelligence, specifically to the goals of accelerating and augmenting emotional intelligence. Some examples of artificial intelligence applications that are investigated include machine learning and deep learning. It offers research on tools that may be utilised by system architects and designers to ease and streamline the process of developing deep learning systems. Specifically, we employ a deep learning approach for labelling feelings in our research. In this method, a huge number of sensor inputs from different modalities are added and subtracted in an iterative process. Our dataset was derived from information collected by multiple smartphones and wearables in a real-world investigation. It does so by displaying signal dynamics and the temporal correlations between the various sensor modalities, all of which are gathered from a unified model that incorporates on-body, ambient, and location sensors. The raw sensor data is fed into numerous learning methods, including a hybrid that employs both convolutional neural networks and long short-term memory recurrent neural networks (CNN-LSTM). Thus, feature engineering and extraction can be carried out automatically.