Choosing good features to represent objects can be crucial to the success of supervised machine learning methods. Recently, there has been a great interest in applying data mining techniques to construct new classification features. The rationale behind this approach is that patterns (feature-value combinations) could capture more underlying semantics than single features. Hence the inclusion of some patterns can improve the classification performance. Currently, most methods adopt a two-phases approach by generating all frequent patterns in the first phase and selecting the discriminative patterns in the second phase. However, this approach has limited success because it is usually very difficult to correctly identify important predictive patterns in a large set of highly correlated frequent patterns In this paper, we introduce the minimal predictive patterns framework to directly mine a compact set of highly predictive patterns. The idea is to integrate pattern mining and feature selection in order to filter out non-informative and redundant patterns while being generated. We propose some pruning techniques to speed up the mining process. Our extensive experimental evaluation on many datasets demonstrates the advantage of our method by outperforming many well known classifiers.
展开▼
