Cost-sensitive decision tree with multiple resource constraints

Measuring an attribute may consume several types of resources. For example, a blood test has a cost and needs to wait for a result. Resource constraints are often imposed on a classification task. In medical diagnosis and marketing campaigns, it is common to have a deadline and budget for finishing the task. The objective of this paper is to develop an algorithm for inducing a classification tree with minimal misclassification cost under multiple resource constraints. To our best knowledge, the problem has not been studied in the literature. To address this problem, we propose an innovative algorithm, namely, the Cost-Sensitive Associative Tree (CAT) algorithm. Essentially, the algorithm first extracts and retains association classification rules from the training data which satisfy resource constraints, and then uses the rules to construct the final decision tree. The approach can ensure that the classification task is done within the specified resource constraints. The experiment results show that the CAT algorithm significantly outperforms the traditional top-down approach and adapts very well to available resources.