Application of Hybrid Deep Learning Algorithm for Sentimental Analysis & Emotional Behavior for Recognition and Classification on Twitter Data Set

Abstract

The rapidity with which technology is developing is changing the way people talk to one another. Face book, Twitter, and Instagram are just a few of the many online communities where people may connect with others who share their interests and express themselves through the written word, visual media, and sound. This opens up the possibility of examining users' emotions and sentiments in their online communications by analyzing data from social networks. In order to glean useful information from user reviews, psychologists utilize a text-processing technique called psychological analysis. Depression is an emotion that can have a significant impact on a person's ability to function normally. Worldwide, the percentage of people dealing with recurrent emotions keeps rising. Self-harming activities are common among depressed people and can sometimes lead to suicide. Those who study the mind rely on social media to spot signs of depression-related behavior and activities. Many indicators of depression's beginnings can be gleaned from a person's social network, including a lack of interest in others, a preoccupation with one's own needs, and increased day- and nighttime activity. High social media use has been linked to more feelings of depression, according to recent research. Recognizing persons with mental health issues and getting them help as soon as feasible is a very difficult endeavor. Patient interviews & PHQ scores were traditionally used to diagnose depression, but these procedures are highly inaccurate. Machine learning, deep learning, & artificial intelligence are examples of cutting-edge technologies that have contributed significantly to these breakthroughs. This study also intends to employ machine learning methods to identify a depressed Twitter user by analyzing their online activity and tweets. To this end, we gathered variables from a user's network activity and tweets and used them to train and evaluate classifiers that can determine whether a user is depressed. The results of this work were tested on datasets taken directly from the scientific literature. For highly accurate depression detection, the suggested Hybrid method outperforms the current gold standard.