Evaluation of a Probabilistic Approach to Classify Incomplete Objects Using Decision Trees

摘要

We describe an approach to fill missing values in decision trees during classification. This approach is derived from the ordered attribute trees method, proposed by Lobo and Numao in 2000, which builds a decision tree for each attribute and uses these trees to fill the missing attribute values. Both our approach and theirs are based on the Mutual Information between the attributes and the class. Our method takes the dependence between attributes into account by using the Mutual Information. The result of the classification process is a probability distribution instead of a single class. In this paper, we present tests performed on some real databases using our approach and Quinlan's method. We analyse the classification results of some instances in test data and finally we discuss some perspectives.