Mining Online Social Networking Data for Detecting Mental Disorders

EasyChair Preprint 5237

Abstract

The growth of social network communication has resulted in risky use. Recently, there has been an increase in the number of social network mental disorders (SNMD), such as reliance on cybernetic relationships, data overload, and network constriction. Currently, these mental illnesses’ manifestations are passively detected, resulting in late clinical intervention. In this paper, the authors argue that mining online social activity provides a way to systematically classify the SNMD at an early level. Since the mental state cannot be detected directly from reports of online social interactions, it is difficult to identify SNMD.Our new and groundbreaking approach to SNMD detection is not focused on self-disclosure of mental factors by psychological questionnaires. Instead, we suggest a machine learning algorithm called SNMD (Detection of Mental Illnesses in Social Networks), which uses features derived from social network data to reliably classify possible SNMD cases. We also propose a new SNMD-based tensor model (STM) to increase accuracy and use multiple sources learning in SNMD. We boost STM’s efficiency with performance guarantees to increase its scalability.A user analysis with a large number of network users is used to test our framework. We evaluate the features of the three types of mental disorders using feature analysis and SNMD in large-scale data sets.

Keyphrases: Decision Tree Classifier, feature extraction, illness identification, social network mental

@booklet{EasyChair:5237,
  author    = {Surabhi Mohite and Satish Gujar},
  title     = {Mining Online Social Networking Data for Detecting Mental Disorders},
  howpublished = {EasyChair Preprint 5237},
  year      = {EasyChair, 2021}}