Rafa, Elean Sugafta and Mahmooda, Adeeba and Sase, Takumi (2024) Suicide risk prediction using artificial intelligence. International Journal on Perceptive and Cognitive Computing (IJPCC), 10 (2). pp. 1-7. E-ISSN 2462-229X
PDF
- Published Version
Restricted to Repository staff only Download (483kB) | Request a copy |
Abstract
Over the past decade, social media has been attracting a growing number of people to the online space. Due to the increase in internet usage, a huge number of text data has been produced. Such data can reflect users’ mental health status, but it is still challenging to predict suicide risk from data, due to the high complexity of texts. This research aims to predict the suicide risk from Reddit posts using artificial intelligence (AI). The data were collected from the Kaggle dataset, which included postings of suicide subreddits. The data were pre-processed through natural language processing techniques. Logistic regression, naive Bayes, and random forest models were then used for classifying the Reddit users, i.e., to predict if they are in a suicidal or non-suicidal mental state. These models were compared to identify an AI approach that provides the best performance among the three models. Then, the logistic regression model with doc2vec showed the highest precision of 0.92, recall 0.92, and F1 score of 0.92.
Item Type: | Article (Journal) |
---|---|
Uncontrolled Keywords: | Suicide Risk, Artificial Intelligence, Machine Learning, Reddit |
Subjects: | Q Science > QA Mathematics > QA75 Electronic computers. Computer science |
Kulliyyahs/Centres/Divisions/Institutes (Can select more than one option. Press CONTROL button): | Kulliyyah of Information and Communication Technology > Department of Computer Science Kulliyyah of Information and Communication Technology > Department of Computer Science Kulliyyah of Information and Communication Technology Kulliyyah of Information and Communication Technology |
Depositing User: | Dr. Takumi Sase |
Date Deposited: | 01 Aug 2024 11:36 |
Last Modified: | 01 Aug 2024 11:36 |
URI: | http://irep.iium.edu.my/id/eprint/113571 |
Actions (login required)
View Item |