Predicting Despair in Bangladeshi Undergraduates Utilizing Machine Studying: Our Journey
Psychological well being consciousness is essential, particularly in creating nations like Bangladesh, the place stigma and restricted sources typically hinder well timed analysis and therapy. As an undergraduate enthusiastic about each expertise and psychology, I led a analysis venture exploring how machine studying might successfully predict despair amongst college college students in Bangladesh.
Despair amongst college college students in Bangladesh is alarmingly excessive, influenced by numerous components reminiscent of tutorial pressures, monetary struggles, relationship points, and household issues. Sadly, many college students don’t obtain the assist they want on account of cultural stigmas and restricted psychological well being sources. Our objective was to bridge this hole, utilizing expertise to establish at-risk college students early.
Our method was methodical:
- Information Assortment: We designed an in depth survey in session with psychologists, counselors, and professors. The survey mixed demographic questions, way of life components, and two standardized scales — Beck Despair Stock-II (BDI-II) and the Despair Anxiousness Stress Scales (DASS 21-Bangla).
- Participant Engagement: Carried out from July to September 2018, our survey reached over 935 undergraduate college students from numerous personal universities throughout Bangladesh.
- Information Cleansing: To make sure knowledge high quality, we meticulously cleaned and validated responses, narrowing our dataset to 577 dependable situations.
We examined three machine studying algorithms — Random Forest (RF), Help Vector Machine (SVM), and Ok-Nearest Neighbors (KNN) — to search out probably the most correct predictor of despair.
Random Forest emerged as one of the best performer, offering roughly 75% accuracy, superior precision, recall, and fewer false negatives in comparison with different fashions.
- Machine studying successfully predicts early-stage despair in Bangladeshi undergraduates.
- Components like tutorial stress, household pressures, and private relationships considerably contribute to despair.
- Well timed detection by way of predictive modeling can doubtlessly cut back extreme outcomes reminiscent of suicide.
Our analysis demonstrates that expertise generally is a highly effective instrument in psychological well being care. By predicting despair early, we hope to allow sooner, focused interventions. Transferring ahead, we goal to broaden our dataset and refine our fashions for even higher accuracy.
Psychological well being shouldn’t be taboo. Expertise can — and will — play a job in supporting college students once they want it most.
For a extra detailed exploration of our research and methodologies, you’ll be able to entry the complete paper right here: Predicting Depression in Bangladeshi Undergraduates using Machine Learning.
Be aware: This analysis was carried out in 2019, and the sector of psychological well being prediction utilizing machine studying has probably advanced since then. Readers are inspired to seek the advice of newer research for the newest developments on this space.
Led by Rezwan Hassan Khan (myself), alongside Ahnaf Atef Choudhury, Nabuat Zaman Nahim, Sadid Rafsun Tulon, Samiul Islam, and Amitabha Chakrabarty.
