A LOF Algorithm-Based Multivariate Process Monitoring Scheme for Mixed-Type Data

doi:10.3969/j.issn.1007-7375.2020.03.019

Abstract

Abstract: A LOF algorithm-based mixed-type data control chart (Mixed-type data Local Outlier Factor Control Chart, MLOF) was proposed to solve the monitoring problem of mixed-type data with ordinal and nominal variables. During the process of detecting changes in process variables, MLOF control chart fully considers the information entropy of nominal categorical variables and the rank of ordinal categorical variables. It measures the abnormality of observation point based on density. The performance of MLOF control chart and the existing mixed-type data control chart on outlier detection was compared using a simulation case based on credit card application data set and a real data case based on German credit card data set. The simulation case included 30 monitoring scenarios. The results show MLOF control chart performs best in 57% of these scenarios. The real data case also verifies that the MLOF control chart is more suitable for the mixed-type data monitoring process with large data volume and complicated clustering situation.

Key words: multivariable mixed-type data, information entropy, distance metric, LOF algorithm, MLOF control chart

CLC Number:

TP277

ZHANG Qiaowei, LI Yanting. A LOF Algorithm-Based Multivariate Process Monitoring Scheme for Mixed-Type Data[J]. Industrial Engineering Journal, 2020, 23(3): 145-153.

References

[1] HOTELLING H. Multivariate quality control//Techniques of statistical analysis[M]//New York: McGraw-Hill, 1947: 111-184.
[2] WOODALL W H, NCUBE M M. Multivariate CUSUM quality-control procedures[J]. Technometrics, 1985, 27(3): 285-292
[3] LOWRY C, WOODALL W, CHAMP C, et al. A multivariate exponentially weighted moving average control chart[J]. Technometrics, 1992, 34(1): 8
[4] BAKIR S. Distribution-free quality control charts based on signed-rank-like statistics[J]. Communications in Statistics, 2006, 35(4): 15
[5] GRAHAM M A, CHAKRABORTI S, HUMAN S W. A nonparametric EWMA sign chart for location based on individual measurements[J]. Quality Engineering, 2011, 23(3): 227-241
[6] ZOU C, WANG Z, TSUNG F. A spatial rank-based multivariate EWMA control chart[J]. Naval Research Logistics (NRL), 2012, 59(2): 91-110
[7] NING X, TSUNG F. A density-based statistical process control scheme for high-dimensional and mixed-type observations[J]. IIE Transactions, 2012, 44(4): 301-311
[8] BREUNIG M M, KRIEGEL H P, NG R T, et al. LOF: identifying density-based local outliers[C/OL]. (2000-05-16). https://doi-org/10.1145/335191.335388.
[9] TUERHONG G, KIM S B. Gower distance-based multivariate control charts for a mixture of continuous and categorical variables[J]. Expert Systems with Applications, 2014, 41(4): 1701-1707
[10] DING D, TSUNG F, LI J. Rank-based process control for mixed-type data[J]. IIE Transactions, 2016, 48(7): 673-683
[11] HUANG Z. Clustering large data sets with mixed numeric and categorical values[C]. Jeju: Springer, 1997: 21-34.
[12] AHMAD A, DEY L. A k-mean clustering algorithm for mixed numeric and categorical data[J]. Data & Knowledge Engineering, 2007, 63(2): 503-527
[13] CHEUNG Y M, JIA H. Categorical-and-numerical-attribute data clustering based on a unified similarity metric without knowing cluster number[J]. Pattern Recognition, 2013, 46(8): 2228-2238
[14] 冯立伟, 张成, 李元, 等. 基于统计模量和局部近邻标准化的局部离群因子故障检测方法[J]. 计算机应用, 2018, 38(4): 965-970
FENG Liwei, ZHANG Cheng, LI Yuan, et al. Local outlier factor fault detection method based on statistical pattern and local nearest neighborhood standardization[J]. Journal of Computer Applications, 2018, 38(4): 965-970
[15] ZHU J, WANG Y, ZHOU D, et al. Batch process modeling and monitoring with local outlier factor[J]. IEEE Transactions on Control Systems Technology, 2018(99): 1-14
[16] JOHNSON R W. An introduction to the Bootstrap[J]. Teaching Statistics, 2001, 23(2): 49-54
[17] DUA D, GRAFF, C. UCI machine learning repository[DB/OL]. [2019-10-16]. http://archive.ics.uci.edu/ml.
[18] SUKCHOTRAT T, KIM S B, TSUNG F. One-class classification-based control charts for multivariate process monitoring[J]. IIE Transactions, 2009, 42(2): 107-120
[19] HE Q P, WANG J. Fault detection using the k-nearest neighbor rule for semiconductor manufacturing processes[J]. IEEE Transactions on Semiconductor Manufacturing, 2007, 20(4): 345-354
[20] FAWCETT T. An introduction to ROC analysis[J]. Pattern Recognition Letters, 2006, 27(8): 861-874