基于决策树的抑郁障碍预测模型

A Depression Prediction Model Based on Decision Tree

从躯体症状、心理状态、社会支持和情绪控制4个方面共26条特征入手,基于C4.5决策树算法,构建了一个抑郁障碍预测模型,预测准确率达到69%以上.实验表明,失眠、自杀意念、躯体疾病等是抑郁的危险因素.对比Apriori算法,可以看出决策树在计算效率、规则提取和解释上有着明显的优势,在一定程度上可以辅助医疗工作者进行抑郁障碍诊断.

From four aspects of physical symptoms,psychological states,social supportand emotional control including 26 depression features as the survey questions,the C4.5 decision tree algorithm was introduced to construct a prediction model for depression and the prediction accuracy was over 69%.The experiments showed that insomnia,suicidal attempt and physical diseases were risk factors of depression.Comparing with Apriori algorithm,it could be seen that the decision tree had obvious advantages in calculation efficiency,rule extraction and interpretation,to a certain degree,this model could help medical professionals to diagnose the potential depressed persons.