A Cost-sensitive Decision Tree Optimized Algorithm Based on Adaptive Mechanism

The problem of cost is increasingly becoming the hotspot of academic research. Therefore, it is necessary to take the cost into account in the research of decision tree algorithm. In this paper, a new algorithm based on adaptive mechanism was proposed. The heuristic function was improved to solve the high cost of the existing cost-sensitive decision tree problem and the multi-valued attribute bias problem. The adaptive determing parameters mechanism, adaptive selecting the cut point mechanism and adaptive removing attribute mechanism were all applied to the process of tree building to improve the model building efficiency. Finally, the improved "probability refusing pruning" strategy was applied to the model of decision tree. Experiments showed that the efficiency of the new algorithm was improved obviously on large data set, and it can also be used to build a model of decision tree with low redundancy and low cost.