Application And Realization Of Data Mining In E-commerce

The purpose and the main task of data mining are: inproves the people's awareness of the large number of seemingly unrelated data, a deeper understanding, and application by finding the useful new laws and new concepts. Data mining technology is the most cutting-edge research direction on the database and the Information Decision Field, and It is also a hot issue of common concern on Academia and the business community.Because the traditional classical algorithm association rules --Apriori may generate a large number of candidate sets, and it has the shortcomings of scanning the database repeatly. The proposed algorithm is L_Apriori .It's based on the style that not access all the databases, but only care about the number of items in the database assumes that the number of items greater than or equal maximum frequent item sets business when it's miniing frequent itemsets. This style can save the storage space, especially for the sparse data.Paper first introduces the current development of data mining technology and the research significance of the subject.To analyze the classical rules of algorithm Apriori algorithm by an example.For the problem, This paper proposes an improved algorithm L_Apriori. By simulation, This algorithm is proved,when the support or data of the database is a value. This algorithm can increase the mining rate. So this algorithm is better than Apriori algorithm.At last,we can design the data mining system based on electronic commerce by use the L_Apriori algorithm. This system has realized the analysis of members and commodities in this system. The system include three functional modules:members of mining module, commodity analysis mining module, forecast Recommend mining module. Member of mining module includes analysis personal information of members, level, the cause of customers lossing. Product of mining module includes sales trend analysis, sales analysis, correlation analysis of goods. Forecast mining module according to members who recommended the purchase of information, To recommend the goods he may interested. Finally, To analyze the mining data in a database, And displayed through the table. Use the Hibernate and Spring, struts2 Excellent frameworks ,and based on J2EE platform. The struts can provide a great convenience on develop the system because it is easy to use and flexible and cross-platform. The system has great practical value.