基于机器学习的大学生自杀风险预测与分析

发布时间：2018-03-09 13:22

本文选题：机器学习　切入点：大学生自杀风险　出处：《现代电子技术》2017年21期 　论文类型：期刊论文

【摘要】：针对当前大学生自杀风险预测方法的预测误差大,精度低等问题,提出基于机器学习的大学生自杀风险预测与分析方法。首先采集大学生自杀风险的影响因素,并结合样本熵对大学生自杀风险的影响因素子序列进行复杂度分析,根据熵值进行归类叠加处理;然后利用交叉验证理论和重构相空间理论建立大学生自杀因素间的相关系数矩阵;最后结合支持向量机理论确定大学生自杀风险预测的决策函数,以此为依据组建大学生自杀风险预测模型。结果表明,该方法的大学生自杀风险分析预测精度高,实用性较强,可以为积极开展自杀预防与危机干预提供重要的理论依据。
[Abstract]:Aiming at the problems of large prediction error and low precision in the current prediction methods of college students' suicide risk, a method of predicting and analyzing college students' suicide risk based on machine learning is put forward. Firstly, the influencing factors of college students' suicide risk are collected. Combined with the sample entropy to analyze the complexity of the influencing factors of suicide risk of college students, according to the entropy value to classify and superposition; Then the correlation coefficient matrix between suicide factors of college students is established by cross-validation theory and phase space reconstruction theory. Finally, the decision function of suicide risk prediction of college students is determined by combining support vector machine theory. The results show that this method has high accuracy and practicability, and can provide an important theoretical basis for active suicide prevention and crisis intervention.
【作者单位】：洛阳理工学院;
【分类号】：B846;TP181

【相似文献】