GMDH-Based Outlier Detection Model in Classification Problems
- PDF / 457,452 Bytes
- 17 Pages / 595.276 x 841.89 pts (A4) Page_size
- 76 Downloads / 152 Views
GMDH-Based Outlier Detection Model in Classification Problems∗ XIE Ling · JIA Yanlin · XIAO Jin · GU Xin · HUANG Jing
DOI: 10.1007/s11424-020-9002-6 Received: 3 January 2019 / Revised: 17 April 2019 c The Editorial Office of JSSC & Springer-Verlag GmbH Germany 2020 Abstract In many practical classification problems, datasets would have a portion of outliers, which could greatly affect the performance of the constructed models. In order to address this issue, we apply the group method of data handing (GMDH) neural network in outlier detection. This study builds a GMDH-based outlier detection (GOD) model. This model first implements feature selection in the training set L using GMDH neural network. Then a new training set L can be obtained by mapping the selected key feature subset. Next, a linear regression model can be constructed in the set L by ordinary least squares estimation. Further, it eliminates a sample from the set L randomly every time, and then rebuilds a linear regression model. Finally, outlier detection is realized by calculating Cook’s distance for each sample. Four different customer classification datasets are used to conduct experiments. Results show that GOD model can effectively eliminate outliers, and compared with the five existing outlier detection models, it generally performs significantly better. This indicates that eliminating outliers can effectively enhance classification accuracy of the trained classification model. Keywords
Classification problem, Cook’s distance, feature selection, GMDH, outlier detection.
XIE Ling School of Medical Information Engineering, Zunyi Medical University, Zunyi 563006, China. Email: xie [email protected]. JIA Yanlin Business School, Sichuan University, Chengdu 610064, China. Email: [email protected]. XIAO Jin (Corresponding author) · GU Xin Business School, Sichuan University, Chengdu 610064, China; Soft Science Institute, Sichuan University, Chengdu 610064, China. Email: [email protected]; [email protected]. HUANG Jing (Corresponding author) School of Public Administration, Sichuan University, Chengdu 610064, China. Email: [email protected]. ∗ This research was partly supported by the Major Project of the National Social Science Foundation of China under Grant No. 18VZL006; the National Natural Science Foundation of China under Grant Nos. 71571126 and 71974139; the Excellent Youth Foundation of Sichuan Province under Grant No. 20JCQN0225; the Tianfu Ten-thousand Talents Program of Sichuan Province; the Excellent Youth Foundation of Sichuan University under Grant No. sksyl201709; the Leading Cultivation Talents Program of Sichuan University; the Teacher and Student Joint Innovation Project of Business School of Sichuan University under Grant No. LH2018011; the 2018 Special Project for Cultivation and Innovation of New Academic, Qian Platform Talent under Grant No. 5772-012. This paper was recommended for publication by Editor WANG Shouyang.
2
1
XIE LING, et al.
Introduction
Classification problem has been one of the key research topics in the field of data min
Data Loading...
