Proceedings of The International Conference on Emerging Trends in Artificial Intelligence and Smart Systems, THEETAS 2022, 16-17 April 2022, Jabalpur, India

Research Article

Identifying the Signature of Suicidality : A Machine Learning Approach

Download302 downloads
  • @INPROCEEDINGS{10.4108/eai.16-4-2022.2318160,
        author={Md Naimur  Rahaman and Sudipto  Chaki and Md Shovon  Biswas and Milon  Biswas and Shamim  Ahmed and Md. Julkar Nayeen  Mahi and Nuruzzaman  Faruqui},
        title={Identifying the Signature of Suicidality : A Machine Learning Approach},
        proceedings={Proceedings of The International Conference on Emerging Trends in Artificial Intelligence and Smart Systems, THEETAS 2022, 16-17 April 2022, Jabalpur, India},
        publisher={EAI},
        proceedings_a={THEETAS},
        year={2022},
        month={6},
        keywords={suicide prediction machine learning svm feature engineering},
        doi={10.4108/eai.16-4-2022.2318160}
    }
    
  • Md Naimur Rahaman
    Sudipto Chaki
    Md Shovon Biswas
    Milon Biswas
    Shamim Ahmed
    Md. Julkar Nayeen Mahi
    Nuruzzaman Faruqui
    Year: 2022
    Identifying the Signature of Suicidality : A Machine Learning Approach
    THEETAS
    EAI
    DOI: 10.4108/eai.16-4-2022.2318160
Md Naimur Rahaman1, Sudipto Chaki1,*, Md Shovon Biswas1, Milon Biswas1, Shamim Ahmed1, Md. Julkar Nayeen Mahi2, Nuruzzaman Faruqui2
  • 1: Bangladesh University of Business and Technology, Dhaka, Bangladesh
  • 2: Daffodil International University, Dhaka, Bangladesh
*Contact email: sudiptochakibd@gmail.com

Abstract

Suicide is a serious problem in today’s world. It would be useful to detect it sooner to lower the death rate. This research proposes a machine learning-based prediction model in this respect. This research uses four traditional machine learning techniques to look at suicidality in various groups of people. To achieve higher levels of accuracy, we have included feature engineering in our suggested model. To accomplish so, a Select-K-Best feature scoring system has been developed. The developed procedure assists in selecting the top 12 scoring characteristics while improving the performance accuracy of the suggested model. In this experimental setup, Random Forest, XGBoost, SVM, and LightGBM prediction algorithms has been used for accuracy prediction. Finally, accuracy levels are measured by fitting our model to the test dataset and found that the XGBoost technique has an accuracy of 89%, which is greater than other implemented ML models.