A peek into one’s emotion through pen: analyzing individual’s emotion through their thoughts expressed on various social media platforms using ensemble classifier

Abstract

Maximum countries of the world are considered as the youth or young nation. But unfortunately, these young ones are the ones suffering the most from emotional distress and suicide ideation. In India alone, every hour one student commits suicide and nearly 28 per day. (NRCB). The NCRB data shows that 10,159 students died by suicide in 2018, an increase from 9,905 in 2017, and 9,478 in 2016 [1]. According to the survey conducted by CSDS in 2017 between the age group of 15-34 approximately one out of every four youth moderately suffered from depression, loneliness, worthlessness, and suicidal thoughts. 6% of them got suicidal thoughts at least once [1]. Despite increasing knowledge, most affected young people do not receive mental health care because the professionals can not easily identify such groups of people [2]. With most people expressing their thoughts and emotions on social media whether positive or negative, examining their posting on the internet has become an essential part of identifying a person’s emotion. Manual identification and examining of this data can be very time-consuming so this paper uses a much better and technical aspect to solve this problem. A dataset is formed by the collection of a few individual’s daily blogs and texts and a combination of machine learning classifiers is used to identify an individual’s emotion. Then the accuracy of these classifiers is tested and the best classifier is used in the system to predict people’s emotions. The result of the analysis shows that the classifier in the system performs better than the other individual classifiers, professionals can understand more people’s emotions and see early signs of emotional distress and provide them with assistance.