Analysis of Suicide Factors in the Chinese Population Based on Non-Negative Matrix Factorization Algorithm

This paper analyzes the factors that contribute to suicide and proposes an innovative suicide tendency prediction model based on the NMF algorithm. Using 52 suicide samples from newspaper reports, assisted by the SVD algorithm, it is found that sui-cide can be better explained by implicit factors instead of explicit ones, and the NMF algorithm is more effective than MLR in sui-cide prediction. The magnitude of the Frobenius Norm of NMF is the quantitative expression to evaluate one's suicide tendency, where a higher Frobenius Norm indicates a lower suicide propensity. The theoretical and practical implications are discussed.