The Eclipse plug-in technology and data mining techniques were introduced. It was analyzed the mainstream of today’s open source data mining platforms, and it was compared, collated and summarized by means of experiments. Based on the experiments, the solution scheme was proposed. It was designed and implemented that an open source data mining platform based on Eclipse plug-in technology. The platform is modular, so it’s easy to maintain and extend. And there is no dependency between its processing unit and the data container, which makes it more suitable to a distributed environment and independent development.

  The decision tree technique is a very effective method for classifying large datasets. By means of constructing a decision tree mode l, the technique picks up valuable classified rules , and helps the decision - makers to make out an exact forecast. The technique has widely applied in a great many fields. The technology is adopted to construct the decision tree model of the edibility of mushroom , which provides scientific basis for differentiating the edibility of mushroom by way of the mushroom property. The calculation of the decision tree uses the C4. 5 algorithm , which takes information gain ratioas attribute choice criterion. The experiment result shows that although the decision tree gives an unbalanced structure ,understandable decision rules are obtained from the decision tree.

    If an enterprise does not monopolize the market , most members o n the market are not its customers.In order to attract the potential customers, enterprises must spend a large amount of time and energy on understanding and analyzing user' s behaviors. Among the traditional commercial affairs , it possibly depends on intuition but no t on facts, because they cannot trace the non -customer' s behavior .But in the ERM data warehouse systems of oriented e -commerce , user' s behaviors can be understood and analyzed through click -stream data .User' s activities can be linked to make a complete access history by using clustering methods. Depending on it , more customers are attracted and market requirements are satisfied better .Through comparing the difference between CRM system and ERM system in B2C environment , the schema of ERM data warehouse was discussed .The schema contains the user -behavior -fact table to record all activities of different users, the basic data dimensional tables and clustering data dimensional tables .An improved method of the schema of ERM data warehouse under the different E-commerce environments was also discussed.