Rules-based Classification with Limited Cost

In test cost-sensitive decision systems, it is difficulty for us to find an optimal attribute set and construct a quality classifier with limited cost. The minimal test cost-sensitive attribute reduction is proposed to address the former problem. However, it is inevitable to remove some good even better attributes in the minimal test cost-sensitive attribute reduction. As a result, the classification accuracy is not high in the minimal test cost attribute reduct. Suppose we have limited cost more than the minimal test cost, we can select addition important attributes to improve the classification accuracy. Therefore, our work includes two aspects. 1) Find an optimal attribute set with limited cost. We put forward a good method to find an attribute set based on the limited cost. 2) We improve the decision trees to build a quality classifier. We construct the root node of the decision tree by several best attribute values. These values just cover the entire dataset. So we can generate more quality rules than ID3. Experimental results indicate that the improved decision tree gets higher accuracy than ID3, and in comparison to the minimal test cost reduct, attribute selection based on the limited cost is feasible to improve the quality of classifiers.