基于Clementine的计算机状态监测研究  

Computer Condition Monitoring Based on Clementine

作　　者：夏日[1] 白雪[1] 

机构地区：[1]安徽财经大学管理科学与工程学院,安徽蚌埠233000

出　　处：《长江大学学报（自科版）（上旬）》2015年第1期44-48,5,共5页JOURNAL OF YANGTZE UNIVERSITY (NATURAL SCIENCE EDITION) SCI ＆ ENG

基　　金：安徽省省级质量工程项目(2012zy037);安徽财经大学精品资源共享课程项目(aczygxkc201415)

摘　　要：数据挖掘技术中的神经网络和C5.0可以用于构建监督和预测的模型。将神经网络和C5.0模型运用于监测计算机的状态信息及识别和预测相关的故障状态。基于Clementine环境,通过采集状态已知情形下的计算机状态数据,作为训练样本来建立神经网络模型和C5.0规则模型,利用另外的实测数据进行检验。结果表明,神经网络的预测正确率是99.64%,而C5.0模型的预测正确率为99.68%,两者预测结果的一致性为99.81%。此外,C5.0模型预测结果的准确性要比神经网络预测结果的准确性高。The neural networks and C5.0 in data mining technology are used to supervise and predict models.The neural networks and C5.0 are used to monitor the state information of computer and identify and predict the related failure state.Based on the circumstance of Clementine,the computer data with known acquisition state are used as training samples to establish the artificial neural network model and the C5.0 rule model,and they are checked out with the other detected data.The result shows that the prediction of correct rate of the neural network is 99.64%,while the correct rate of C5.0 model is 99.68%,the consistency of the both two is 99.81%.Therefore,accuracy of C5.0 model is higher than that of the neural network.

关 键 词：CLEMENTINE C5.0模型 神经网络 故障监测 数据挖掘 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]