Volume 31, Issue 1, 2022
DOI: 10.24205/03276716.2022.6005
Depression Prediction in Social Networks Using Whale Optimization Algorithm based Convolution Neural Networks (WOA-CNN)
Abstract
Depression is a major problem that has many different effects on individuals. Many treatments are available to help those who are depressed, but the issue is predicting those who aren't even aware that they are depressed. Furthermore, it is a leading cause of suicide ideation and significantly impairs everyday functioning. Emotion artificial intelligence is a subject of continuing study in emotion detection, particularly in the areas of social media text mining. Datasets derived from social networks are useful in a variety of subjects, including sociology and psychology. However, technical assistance is insufficient, and particular techniques are needed immediately. It encourages the development of deep learning systems for psychology to identify depressed users on social networking platforms. This paper's objective is to propose a sentiment analysis approach based on terminology and man-made criteria for calculating depressive inclination. Next, a depression detection model is built using the Whale Optimization Algorithm based Convolution Neural Networks (WOA-CNN), which takes into account parameter optimization by WOA. WOA algorithm is introduced to optimize the hyperparameters for training a CNN classifier and their layers capably. In ending, the proposed method honors social network users who have developed high-quality online mental health monitoring solutions. The results of proposed WOA-CNN classifier and existing classifiers are measured regarding accuracy, F-measure, recall, precision. Proposed technique outperforms than the state-of-the-art techniques.
Keywords
Convolutional Neural Network (CNN), Whale Optimization Algorithm (WOA), Depression, Preprocessing, Optimization, and Social Media