Organization of Case Base Represented with XML under Quotient Space

General data mining systems rely on experts to give the algorithm to operate data mining according to actual application, people who are not familiar with data mining algorithms canpsilat use these data mining systems, so we put forward data mining system based on Case-Based Reasoning. In this system, XML was used to represent cases. The case base was organized based on fuzzy quotient space model instead of traditional case index tree, which can improve retrieval efficiency and accurate rate at the same time. Experiment indicates the average effect was improved by more than 50 percent, it will become higher if the case base is bigger.