Time-constrained cost-sensitive decision tree induction

摘要: Cost-sensitive decision tree induction is to build a from training data with minimal cost.No previous research has studied how induce the cost if classification task be completed in limited time.This paper proposed an algorithm time-constrained tree.The experiment results show performance of our very satisfactory under different time constraints. A cost-sensitive induced for purpose building that minimizes sum misclassification and test cost. Although this problem been investigated extensively, no study specifically focused on can must within time. Accordingly, we developed generate minimal-cost tree. The main idea behind select attribute brings maximal benefit when sufficient, most time-efficient (i.e., provides per unit time) limited. Our experimental highly various constraints across distinct datasets.