Active learning using pessimistic expectation estimators

• Autorzy: Rokach L. , Naamani L. , Shmilovici A.
• Języki publikacji: EN

Abstrakty

EN

Active learning is the process in which unlabeled instances are dynamically selected for expert labelling, and then a classifier is trained on the labeled data. Active learning is particularly useful when there is a large set of unlabeled instances, and acquiring a label is costly. In business scenarios such as direct marketing, active learning can be used to indicate which customer to approach such that the potential benefit from the approached customer can cover the cost of approach. This paper presents a new algorithm for cost-sensitive active learning using a conditional expectation estimator. The new estimator focuses on acquisitions that are likely to improve the profit. Moreover, we investigate simulated annealing techniques to combine exploration with exploitation in the classifier construction. Using five evaluation metrics, we evaluated the algorithm on four benchmark datasets. The results demonstrate the superiority of the proposed method compared to other algorithms.

Słowa kluczowe

EN

cost-sensitive learning; active learning; direct marketing; decision tree

• Wydawca: Systems Research Institute, Polish Academy of Sciences
• Czasopismo: Control and Cybernetics
• Rocznik: 2009
• Tom: Vol. 38, no 1
• Strony: 261--280
• Opis fizyczny: Bibliogr. 25 poz., wykr.

Twórcy

autor

Rokach L.

autor

Naamani L.

autor

Shmilovici A.

• Department of Information System Engineering, Ben-Gurion University of the Negev, P.O.Box 653, Beer-Sheva 84105, Israel

Bibliografia

• BLAKE, C.L. and MERZ, C.J. (1998) UCI Repository of machine learning databases. Irvine, CA: University of California, Department of Information and Computer Science, http://www.ics.uci.edu/~mlearn/MLReposi-tory.html.
• BROWN, L.D., CAI, T.T. and DASGUPTA, A. (2001) Interval Estimation for a Binomial Proportion. Statistical Science 16 (2), 101-117.
• COHN, D.A., GHAHRAMANI, Z. and JORDAN, M.I. (1996) Active learning with statistical models. Journal of Artificial Intelligence Research 4, 129-145.
• DEMSAR, J. (2006) Statistical Comparisons of Classifiers over Multiple Data Sets. Journal of Machine Learning Research 7, 1-30.
• ELKAN, C. (2001) The foundations of cost-sensitive learning. Proceedings of the 17th International Joint Conference on Artificial Intelligence. Morgan Kaufmann, 973-978.
• HEDAYAT, A.S, SLOANE, N.J.A and STUFKEN, J. (1999) Orthogonal Arrays: Theory and Applications. Springer-Verlag, NY.
• HOLLMEN, J., SKUBACZ, M. and TANIGUCHI, M. (2000) Input dependent misclassification costs for cost-sensitive classifiers. In: Proceedings of the Second International Conference on Data Mining. WIT Press, 495-503.
• KIRKPATRICK, S., GELATT, C.D. and VECCHI, M.P. (1983) Optimization by Simulated Annealing. Science 220 (4598), 671-680.
• LEWIS, D. and GALE, W. (1994) A sequential algorithm for training text classifiers. Proceedings of the International ACM-SIGIR Conference on Research and Development in Information Retrieval ACM Press, 3-12.
• MARGINEANTU, D. (2005) Active Cost-Sensitive Learning. Proceedings of the Nineteenth International Joint Conference on Artificial Intelligence, IJCAI-05. Professional Book Center, 1622-1631.
• MAYER, U.F. and SARKISSIAN, A. (2003) Experimental design for solicitation campaigns. Proceedings of the 9th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, Washington, D.C., USA. ACM Press, New York, NY, USA, 717-722.
• NAAMANI, L. (2008) Cost Sensitive Active Learning for Target Marketing. M.Sc. dissertation. Dept. of Information Systems Engineering, Ben-Gurion University, Israel.
• PUTTEN, P. and SOMEREN, M. (2000) CoIL Challenge 2000: The Insurance Company Case. Published by Sentient Machine Research, Amsterdam. Also a Leiden Institute of Advanced Computer Science Technical Report 2000-09, June 22.
• QUINLAN, J.R. (1993) C4-5: Programs for Machine Learning. Morgan Kaufmann.
• ROKACH, L. (2008) Genetic algorithm-based feature set partitioning for classification problems. Pattern Recognition 41 (5), 1676-1700.
• ROKACH, L. and MAIMON, O. (2005) Top Down Induction of Decision Trees Classifiers: A Survey. IEEE SMC Transactions Part C, 35 (4), 476-487.
• ROKACH, L. and MAIMON, O. (2008) Data Mining with Decision Trees: Theory and Applications. World Scientific Publishing Company.
• ROKACH, L., NAAMANI, L. and SHMILOVICI, A. (2007) Active Learning Using Conditional Expectation Estimators. In: T. Morzy, M. Morzy and Nanopoulos A., eds., Proceeding of the 3rd ADBIS workshop on Data Mining and Knowledge Discovery ADMKD’2007, Varna, Bulgaria. Springer, 83-95.
• ROKACH, L., NAAMANI, L., and SHMILOVICI, A. (2008) Pessimistic Cost-sensitive Active Learning of Decision Trees for Profit Maximizing Targeting Campaigns. Data Mining and Knowledge Discovery 17 (2), 283-316.
• ROY, N. and MCCALLUM, A. (2001) Toward optimal active learning through sampling estimation of error reduction. Proceedings of the International Conference on Machine Learning, San Francisco, CA. Morgan Kaufmann, 441-448.
• SAAR-TSECHANSKY, M. and PROVOST, F. (2007) Decision-Centric Active Learning of Binary-Outcome Models. Information Systems Research 18 (1), 4-22.
• TONG, S. and ROLLER, D. (2000) Support vector machine active learning with applications to text classification. Proceedings of the 17th International Conference on Machine Learning, ICML-2000, July 2, Stanford, CA. Morgan Kaufmann, 999-1006.
• TURNEY, P. (2000) Types of Cost in Inductive Concept Learning. Proceedings of the Cost-Sensitive Learning Workshop at the 17th International Conference on Machine Learning, ICML-2000, July 2, Stanford, CA. Morgan Kaufmann, 60-66.
• WEISS, G. M. and TIAN, YE (2006) Maximizing classifier utility when training information is costly. SIKDD Exploration 8 (2), 31-38.
• ZADROZNY, B. (2005) One-Benefit Learning: Cost-Sensitive Learning with Restricted Cost Information. In: Proc. of the Workshop on Utility-Based Data Mining at the 11th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. ACM Press, New York, 53-58.