Fraud Detection Model Based on the Discovery Symbolic Classification Rules Extracted from a Neural Network
This paper presents a fraud detection model using data mining techniques such as neural networks and symbolic extraction of classification rules from trained neural network. The neural network is first trained to achieve an accuracy rate, the activation of the values in the hidden layers of the neural network is analyzed and from this analysis are generated classification rules. The proposed approach was tested on a set of data from a Colombian organization for the sending and payment of remittances, in order to identify patterns associated with fraud detection. Similarly the results of the techniques used in the model were compared with other mining techniques such as Decision Trees and Naive Bayes. A prototype software was developed to test the model, which was integrated into RapidMiner tool, which can be used as a tool for academic software.
KeywordsData mining neural networks rule extraction detection of fraud experts
Unable to display preview. Download preview PDF.
- 1.Kou, Y., Lu, C.T., Sirwongwattana, S., Huang, Y.P.: Survey of fraud detection techniques. In: Proceedings of IEEE Intl Conference on Networking, Sensing and Control (2004)Google Scholar
- 2.Shen, A., Tong, R., Deng, Y.: Application of classification models on credit card fraud detection. In: International Conference on Service Systems and Service Management, pp. 1–4 (2007)Google Scholar
- 3.Shao, H., Zhao, H., Chang, G.: Applying data mining to detect fraud behavior in customs declaration, vol. 3 (2002)Google Scholar
- 7.Craven, M.W.: Extracting comprehensible models from trained neural networks. PhD thesis, University of Wisconsin -Madison (1996)Google Scholar
- 9.Mayer, H., Huber, Rohde, Tamme: Rule extraction from artificial neural networks. University Salzburg (October 12, 2006)Google Scholar
- 10.Towell, G., Shavlik, J., Noordewier, M.: Refinement of approximate domain theories by knowledge-based neural networks, pp. 861–866 (1990)Google Scholar
- 11.Thrun, S.B.: Extracting symbolic knowledge from artificial neural networks. Revised Version of Technical Research Report I-University Bonn (1994)Google Scholar
- 12.Mcmillan, C., Mozer, M.C., Smolensky, P.: Rule induction through integrated symbolic and subsymbolic processing. In: Advances in Neural Information Processing Systems, pp. 969–976. Morgan Kaufmann, San Francisco (1992)Google Scholar
- 13.Andrés Ramírez, J.G.E.L., Santamaría, W.: Análisis del proceso de giros internacionales y control de lavado de activos mediante minería de datos. Revista Tendencias en Ingeniería de Software e Inteligencia Artificial 2 (2008)Google Scholar
- 14.Oracle, Database sql reference-ora hash (2003), Online http://download.oracle.com/docs
- 15.Rapidminer.: Rapidminer- free library to work data mining (2002)Google Scholar
- 16.Quinlan, J.R.: Induction of decision trees. Machine learning 1(1), 81–106 (1986)Google Scholar