基于C5.0改进算法的焊接工艺参数选择决策树数据挖掘模型及其应用  

The Decision Tree Data Mining Model for Welding Parameters Selection Based on C5.0 Improved Algorithm and Its Application

作　　者：赵丽 程铁信[1] 莫莹 李奇[1] 刘铭杰 

机构地区：[1]天津工业大学管理学院,天津300387

出　　处：《中国管理科学》2016年第S1期177-182,共6页Chinese Journal of Management Science

基　　金：国家自然科学基金青年基金资助项目(71301117)

摘　　要：焊接变形对焊接工艺参数的选择依赖性很强,在影响焊接的各种因素中占主导地位。本文在经典ID3算法的基础上,建立了决策树数据挖掘模型并给出了决策树的C5.0改进算法,并将该模型应用于焊接工艺参数选择。针对具有代表性的压力容器焊接工艺设计过程,综合分析各种工艺参数对焊接变形的影响,挖掘出能够提升焊件质量的焊接工艺参数,降低焊接缺陷。然后,将该模型与神经网络、Logistic回归模型比较,结果显示该模型和算法具有较高准确度,可以对压力容器焊接质控提供很好的决策支持。Welding deformation strongly depends on welding parameters selection, which can affect the welding quality. The decision tree data mining model is established for welding parameters selection and the C5. 0 improved algorithm for the model is given on the basis of the classic ID3 algorithm. With the model,the typical pressure vessel welding process is analyzed,and the welding process parameters are correctly selected, which can improve the quality of welding and reduce welding defects. Then, compared with neural networks,Logistic regression model,the model above is llustrated with higher accuracy,which can provide the decision making support for pressure vessel welding quality control.

关 键 词：数据挖掘 压力容器 焊接 质量控制 

分 类 号：TG40[金属学及工艺—焊接] TP311.13[自动化与计算机技术—计算机软件与理论]