摘要
在盾构机的作业过程中,由于采集设备的不稳定性或被干扰等原因,采集到的数据往往是不完整的,缺失值的存在严重影响了数据挖掘的质量。此外,不同类别的工程数据在空间中分布复杂,难以通过硬聚类进行确切划分。针对此问题,采用基于TS模型的数据填补算法对不完整盾构机数据进行填补,算法首先通过局部距离策略对不完整数据建立模糊聚类模型,将数据集划分为若干个模糊子集,以获取前提参数。然后,采用交替学习策略对填补值和结论参数进行协同式学习。最后,为各个不完整属性列进行建模估计数据缺失值。试验结果表明,无论是在对不完整数据聚类效果还是填补精度上,该方法都具有良好表现。
When the tunnel-boring machine(TBM)is in operation,due to instability or interference of the data-acquiring equipment,the collected data is often incomplete,and those missing values seriously affect quality of data mining.In addition,the distribution of different types of engineering data is complex in space,and it is difficult to be accurately divided by means of hard clustering.In order to solve this problem,the data-imputation algorithm is used to fill incomplete TBM data based on the TS model.Firstly,with the help of this algorithm,a fuzzy-clustering model of incomplete data is set up by means of the local-distance strategy,and the data set is divided into several fuzzy subsets,so as to obtain the premise parameters.Then,the alternate-learning strategy is adopted for collaborative learning of the filling value and the conclusion parameter.Finally,the missing data values are estimated for each incomplete attribute column.The experimental results have verified that this method has desirable performance both in the clustering effect of incomplete data and the imputation accuracy.
作者
王一棠
庞勇
张立勇
孙伟
宋学官
WANG Yi-tang;PANG Yong;ZHANG Li-yong;SUN Wei;SONG Xue-guan(School of Mechanical Engineering,Dalian University of Technology,Dalian 116024;Faculty of Electronic and Electrical Engineering,Dalian University of Technology,Dalian 116024)
出处
《机械设计》
CSCD
北大核心
2022年第3期26-31,共6页
Journal of Machine Design
基金
国家重点研发计划(2018YFB1702502)。
