Learning with Actionable Attributes: Attention -- Boundary Cases!
作者： Indre Zliobaite , Mykola Pechenizkiy
关键词:
摘要: Traditional supervised learning assumes that instances are described by observable attributes. The goal is to learn predict the labels for unseen instances. In many real world applications values of some attributes not only observable, but can be proactively chosen a decision maker. Furthermore, in such maker interested generate accurate predictions, maximize probability desired outcome. For example, direct marketing manager choose color an envelope (actionable attribute), which offer sent client, hoping right choice will result positive response with higher probability. We study how value actionable attribute order outcome settings. emphasize all equally sensitive change actions. Accurate action essential those instances, on borderline (e.g. do have strong opinion). formulate three approaches select at instance level. focus process cases. potential underlying ideas demonstrated synthetic examples and case dataset.
acm.org
narcis.nl参考文章(12)
Jacques Wainer, Hector Geffner, Modeling Action, Knowledge and Control. european conference on artificial intelligence. pp. 532- 536 ,(1998)
Shengchun Deng, Zengyou He, Xiaofei Xu, Data Mining for Actionable Knowledge: A Survey arXiv: Databases. ,(2005)
Zbigniew W. Ras, Alicja Wieczorkowska, Action-Rules: How to Increase Profit of a Company european conference on principles of data mining and knowledge discovery. pp. 587- 592 ,(2000) , 10.1007/3-540-45372-5_70
Mingkun Li, Ishwar K. Sethi, Confidence-based classifier design Pattern Recognition. ,vol. 39, pp. 1230- 1240 ,(2006) , 10.1016/J.PATCOG.2006.01.010
Z HE, X XU, S DENG, R MA, Mining action rules from scratch Expert Systems With Applications. ,vol. 29, pp. 691- 699 ,(2005) , 10.1016/J.ESWA.2005.04.031
Y. Freund, Boosting a weak learning algorithm by majority Information & Computation. ,vol. 121, pp. 256- 285 ,(1995) , 10.1006/INCO.1995.1136
Huaifeng Zhang, Yanchang Zhao, Longbing Cao, Chengqi Zhang, Combined association rule mining knowledge discovery and data mining. pp. 1069- 1074 ,(2008) , 10.1007/978-3-540-68125-0_115
Peter L. Bartlett, Marten H. Wegkamp, Classification with a Reject Option using a Hinge Loss Journal of Machine Learning Research. ,vol. 9, pp. 1823- 1840 ,(2008) , 10.5555/1390681.1442792
C. Ling, Qiang Yang, Rong Pan, Jie Yin, Extracting Actionable Knowledge from Decision Trees IEEE Transactions on Knowledge and Data Engineering. ,vol. 19, pp. 43- 56 ,(2007) , 10.1109/TKDE.2007.10
Zbigniew W. Ras, Elzbieta Wyrzykowska, Li-Shiang Tsay, Action Rules Mining. Encyclopedia of Data Warehousing and Mining. pp. 1- 5 ,(2009)