Minimal Test Cost Feature Selection with Positive Region Constraint

Test cost is often required to obtain feature values of an object. When this issue is involved, people are often interested in schemes minimizing it. In many data mining applications, due to economic, technological and legal reasons, it is neither possible nor necessary to obtain a classifier with 100% accuracy. There may be an industrial standard to indicate the accuracy of the classification. In this paper, we consider such a situation and propose a new constraint satisfaction problem to address it. The constraint is expressed by the positive region; whereas the objective is to minimize the total test cost. The new problem is essentially a dual of the test cost constraint attribute reduction problem, which has been addressed recently. We propose a heuristic algorithm based on the information gain, the test cost, and a user specified parameter λ to deal with the new problem. Experimental results indicate the rational setting of λ is different among datasets, and the algorithm is especially stable when the test cost is subject to the Pareto distribution.