摘要
针对传统端子缺陷检测方法太过落后的问题,提出将随机森林模型应用于端子缺陷检测方面:首先将原始电压数据进行预处理,消除原始数据集中高频噪声和抖动,从而获得初始数据集;通过决策树算法建立单个基模型,再经由多个基模型的多数投票准则最终得到随机森林模型;最后再将模型预测效果与其它机器学习模型在同等数据集的基础上进行比对实验。随机森林模型在端子缺陷检测中准确率为94%,查准率达到了95%,实验结果也表明该模型相比其它分类算法有更好的性能表现,且其预测效果也较为稳定,是更为优越的端子缺陷检测算法。
In order to solve the problem that traditional terminal defect detection methods are too backward,it proposed to apply the random forest model to terminal defect detection in this paper.The initial data set was obtained by preprocessing the original voltage data and eliminating the high frequency noise and jitter in the original data set.A single base model was established by decision tree algorithm,and then the majority voting criteria of multiple base models were used to obtain random forest model.Finally,the prediction effect of the model was compared with other machine learning models on the basis of the same data set.The accuracy of random forest model in terminal defect detection is 94%,and the precision is 95%.Experimental results also showed that this model has better performance than other classification algorithms,and its prediction effect is more stable,and is a superior terminal defect detection algorithm.
作者
陈晨
张捍东
吴玉秀
CHEN Chen;ZHANG Han-dong;WU Yu-xiu(School of Electrical and Information Engineering,Anhui University of Technology,Ma’anshan,243002,Anhui)
出处
《蚌埠学院学报》
2021年第2期97-102,共6页
Journal of Bengbu University
基金
安徽省自然科学基金项目(1608085QF155)。
关键词
端子
决策树
随机森林
机器学习
terminal
decision tree
random forest
machine learning
