A genetic algorithm to attribute reduction with test cost constraint

In many machine learning applications, we need to pay test cost for each data item. Due to limited money and/or time, we also have a constraint on the total test cost. This issue have been recently formalized as the optimal sub-reduct with test cost constraint problem. An information gain based heuristic algorithm has been proposed to deal with it. In this paper, we propose a genetic algorithm which takes advantages of both the test cost information and the search potential of GA. Experimental results on four UCI datasets indicate that the new algorithm generally produces better results than the existing one.