不同行業別虛設行號預警模型之初探

Research on Detecting Model for Fictitious Company of Different Trades

陳祝美(Chu-Mei Chen)

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)

摘要

本研究嘗試運用資料探勘技術於龐大稅務資料，將資料分為15個行業別，利用SPSS Clementine 12.0套裝軟體，進行資料探勘的訓練、測試、模型的建置，並對虛設行號的特徵進行預測，比較不同行業別虛設行號之預警警示效果，俾利找出最佳的預警模式及規則，以改善電腦選案正確率進而提升查核稽徵效率。本研究結果顯示各行業別均以C5.0建置之模型，預測的選案正確率最高，以Discriminant之預測效果最差，其餘各行業別在不同模型各有不同的正確率；另應用不同的建模方式，在關鍵因子不變的情況下，對不同行業別的虛設行號，會有不一樣的選案正確率，顯示不同行業別可能有不同的虛設型態，可以做為實務上選案的參考。

關鍵字

虛設行號 ； 資料探勘 ； 行業別

並列摘要

This research used data mining technique among a huge tax database. The SPSS Clementine 12.0 software was used on training, testing and model-building the data mining process that was looking into 15 different occupations. Yet, I built virtual companies and set up a taxation pre-warning systems for each of them. As the simulation ran, I could find out the best pre-warning systems and their rules. According to the simulation results, computer could analytically tell me which model worked the best on efficient taxation enhancement. At the end, this research showed C5.0 model gave the best result for every occupation taxation performance. On the other end, the discriminant method gave the worst prediction result. However, among the rest of the models I tested, it carried out different accuracies respectively. And yet, if the key factors remained fixed, different models also resulted different accuracies as well. To sum up, modeling simulations on different occupations can be used for reference in real-life practices of case selection.

並列關鍵字

Fictitious Company ； Data Mining ； Trades

參考文獻

[11]顏秀珍、李御璽、王秋光，「改善不平衡資料集中少數類別資料之分類正確性的方法」，電子商務學報，第十一卷第四期， 847-858頁，2009年12月。

[6]賴志鏹，「虛設行號案件調查蒐證之研究」，中原大學財經法律研究所碩士論文，2005。

[16]Shaw M. J., Subramaniam C. and Tan G. W, “Knowledge management and data mining for marketing,”Decision Support Systems, 2001. pp. 127-137.

[17] Fayyad U., Piatetsky G., and Smyth P, “From data mining to knowledge discovery in databases,” AI Magazine, FALL,1996. pp. 37-54.

[19] Grupe, F. H. and Owrang, M. M., “Data Base Mining Discovering New Knowledge and Coperative Advantage,” Information Systems Management, Vol.12, No. 4. Fall,1995, pp.35-41.