運用資料探勘技術提升稽徵機關選案作業之捕捉率

Use Data Mining Techniques to Assist the Tax Authorities in Case Election

王達暐(Ta-Wei Wang)

指導教授 ： 盧以詮

元智大學/資訊學院/資訊管理學系/碩士(2011年)

https://doi.org/10.6838/YZU.2011.00199

摘要

本篇論文主要係探討資料探勘技術是否可用於稽徵機關違章案件之選查工作，以提升人工選案作業可能發生的人為操縱或偏差，經以虛設行號案件進行實證研究，資料探勘分類模型的回應率及捕捉率均高於人工選案，本項資訊技術在導入稽徵機關之稅務行政工作後，咸信可以大幅提升稽徵人力的運用效率，惟本研究認為，對於技術的導入雖屬可行，但建議應配合建立稅務資料探勘選案模型管理系統及標準作業流程，俾利本項作業之順遂。

關鍵字

資料探勘 ； 營業稅 ； 虛設行號

並列摘要

This thesis mainly studied whether Taxation agency could introduce
data mining techniques to assist the process of selecting cases which have violated tax policies and regulations. The results showed that precision and recall rate produced by data mining classification models are better than the traditional cases selection process in the detection of bogus business entity. Consequently, it is believed that Taxation agency will highly improve the efficiency of the allocation of manpower if data mining techniques is introduced into the tax administration activities. However, the study also suggests that the key to facilitate the activities of selecting cases properly is to establish a mechanism and a standard operation procedure for data mining process.

並列關鍵字

Data mining ； Business Tax ； Bogus Business Entity

參考文獻

[8]黃傳益，”使用資料探勘技術於稅務資料中判斷虛設行號之研究”，國立成功大學工程研究所碩士論文，2008。

[14]Sotiris Kotsiantis, Dimitris Kanellopoulos, Panayiotis Pintelas,” Handling imbalanced datasets” , pp5~16 , 2006.

[6] Cecile Levasseur , ”CLASSIFYING NON-GAUSSIAN AND MIXED DATA SETS IN THEIR NATURAL PARAMETER SPACE” , pp10~16 , 2007.

[11] Miin-Shen Yang, Pei-Yuan Hwang, De-Hua Chen,” Fuzzy clusteringalg orithms for mixed feature variables” ,pp2~15 , 2004.

[13]Chung-Chian Hsu et al.,” Extended Naive Bayes classifier for mixed data” , pp5~21 ,2008.