Automatic data mining cross tables with dominate cells using MPT models
Abstract
The main purpose of cross tables with dominate cells data mining is to reveal the latent interactions/factors to explain the causes of these abnormal cells. Based on Multinomial Processing Tree (MPT) Models, this paper presented an automatic data mining approach for analyzing the dominate cells by 1) using a special MPT structure with latent classes to represent the corresponding cross table uniquely according to its own characteristics, and 2) building a set of algorithms including the category classification and hypothesis generation requirement to ensure the entire processes of data mining can be processed automatically. Compared with traditional methods, the proposed approach not only could acquire the quantized latent interactions for interpreting the dominate cells, but also could improve the efficiency and coverage of data mining processes. ©2010 IEEE.
Publication Title
Proceedings - 2010 IEEE International Conference on Intelligent Computing and Intelligent Systems, ICIS 2010
Recommended Citation
You, Y., Qi, H., & Hu, X. (2010). Automatic data mining cross tables with dominate cells using MPT models. Proceedings - 2010 IEEE International Conference on Intelligent Computing and Intelligent Systems, ICIS 2010, 2, 588-591. https://doi.org/10.1109/ICICISYS.2010.5658407
