Designing an algorithm to detect depression in users: A quantitative correlational study.
Although the diagnosis of mental disorders like depression has improved over the last decade, many cases continue to go undetected. The symptoms are often observable on social media platforms. This study seeks to address this issue by designing a program to predict the likelihood and severity of dep...
Saved in:
| Main Author: | |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Royal St. George's College
2021-08-01
|
| Series: | The Young Researcher |
| Subjects: | |
| Online Access: | http://www.theyoungresearcher.com/papers/hodan.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Although the diagnosis of mental disorders like depression has improved over the last decade, many cases continue to go undetected. The symptoms are often observable on social media platforms. This study seeks to address this issue by designing a program to predict the likelihood and severity of depression in users by analyzing their Twitter and iMessage histories. This study employs correlational research to develop and test the program. This research used two phases to measure the program’s accuracy on different sources of data. In phase one, the program was tested on a sample of 2,741 Reddit posts and comments. The program achieved an accuracy rate of 94.80% across several subreddits. Suicidal behaviour/ideation and a depressed mood had the strongest correlations with depression (coefficients of 0.654997 and 0.58218, respectively). Phase two involved testing the program on 6 grade 12 students. During this phase, the program had a recall, precision, and F1 of 1.0, 0.5, and 0.67, respectively. The results suggest that ERDS should integrate data from platforms like Twitter, iMessage, and Reddit because they reflect users’ mental states. |
|---|---|
| ISSN: | 2560-9823 2560-9823 |