Research Of Decision Tree Algorithm In Data Mining And Application In Real Estate Agency

With the quick development of the database technique and the wide uses of DBMS(Database Management System) in the field of real estate, each department is induced to accumulate more and more data. Although many functions can be done efficiently with the current database technique, for example, do query or do statistic, the relationship and rules among data still can't be found. For the real estate agency (one of major operations in the field of real estate), It is a fact problem, which is value of research, how to mine the useful knowledge with huge data, and how to make rapidly the predictions which find out potential buyer and corresponding dealer.As for the problem above, the paper in-depth researches one of the core algorithms which are applied to Classification—decision tree algorithm. According to comparation, ID3 algorithm is adopted to resolve this problem. In the application, with Information Gain of Attribute-value Pairs in Two Levels optimization algorithm, the problem is solved that it is easy to select those attributes whose values is more, while are not alaways the best in ID3 algorithm. Moreover, with the same instance set (the same record set of house information), the decision trees are built separately with ID3 algorithm and mentioned above optimization algorithm, and corresponding rules like that houses of some type are apt to make a deal are created.It is concluded by experiments, the decision tree built with Information Gain of Attribute-value in Two Levels algorithm is more similar to ideal decision tree, the rate of accuracy is higher than one built by ID3 algorithm. If this algorithm is applied to analyse house information in the field of real estate agency,a large amount of house information can be classificated more correctly by the characters of house information so that we can realize the reasonable prediction whether a trade can be successful, whichsupervises more scientifically the real estate enterprises to find out"potential client"of real estate agency.