摘要
构建基于数据挖掘技术的心理障碍预测模型,有效预测心理障碍,提升预测能力。依据随机森林原理,通过投票或计算平均数的决策方式重组随机森林生成的分类树,Bagging算法使用Bootstrap从心理障碍数据样本集内反复抽取子心理障碍数据集,构建分类树模型,通过计算心理障碍样本集的信息增益、信息率等确定分类树分裂节点并构建决策树分支与叶节点,经过剪枝处理后,通过提升随机森林的收敛性、分类能效与相关度,缩小其泛化误差,实现心理障碍预测。实验结果表明:该模型灵敏度达到0.95,预测价值较高;预测准确率达到0.98,且召回率较高,预测能力强。
This paper builds a prediction model of psychological disorders based on data mining technology to effectively predict psychological disorders and improve the prediction ability.According to the principle of bootstrap,a random decision tree is constructed to calculate the average number of decision-making nodes.After pruning,the generalization error is reduced by improving the convergence,classification efficiency and correlation of random forest,and the prediction of psychological disorders is realized.The experimental results show that the sensitivity of the model is 0.95,and the prediction value is high;the prediction accuracy is 0.98,and the recall rate is high,and the prediction ability is strong.
作者
李永明
LI Yongming(Teachers College, Weinan Vocational and Technical College, Weinan 714026, China)
出处
《微型电脑应用》
2021年第11期161-164,共4页
Microcomputer Applications
关键词
数据挖掘
随机森林
决策树
心理障碍
预测模型
泛化误差
data mining
random forest
decision tree
psychological disorder
prediction model
generalization error