Classify Uncertain Data With Decision Tree

This demo presents a decision tree based classification system for uncertain data. Decision tree is a commonly used data classification technique. Tree learning algorithms can generate decision tree models from a training data set. When working on uncertain data or probabilistic data, the learning and prediction algorithms need handle the uncertainty cautiously, or else the decision tree could be unreliable and prediction results may be wrong. In this demo, we will present DTU, a new decision tree based classification and prediction system for uncertain data. This tool uses new measures for constructing, pruning and optimizing decision tree. These new measures arc computed considering uncertain data probability distribution functions. Based on the new measures, the optimal splitting attributes and splitting values can be identified and used in the decision tree. We will show in this demo that DTU can process various types of uncertainties and it has satisfactory classification performance even when data is highly uncertain.