Towards Interpretable Suicide Risk Prediction: A Hybrid Approach with Feature Extraction and Sequential Binary Classification

This paper presents our contribution to the IEEE BigData 2025 Cup Challenge. The goal of the challenge is to develop a classifier capable of predicting varying suicide risk levels based on post sequences. To ensure the interpretability of our model, we adopted an interdisciplinary perspective, examining a range of indicators and their feature subsets that have been shown to be relevant to suicide risk. The model we developed is based on a sequential binary classification structure, integrating the selected features for predicting suicide risk levels. Our proposed framework achieved a weighted F1 score of 0.43 on the validation data, as reported on the leaderboard, and 0.44 on the final test data.