Research And Implementation Of Data Mining Algorithms Based On Logistics Information System

With the fast development of computer technology and internet technology,society informatization and business intelligence grow more and more fiercely. Andeconomic globalization and integration now is the target of modern logistics industry,making logistics enterprises to make full use of information technology to improvetheir market competitiveness and profit. When mining potential information whichmay be decision supportive from enterprise-level huge data, related techniques isnecessary, and data mining technology becomes very important.There are several different data mining technologies, and association rule miningis the most widely used one. Association data mining technology can be applied inCRM, short for Client Relationship Management, ERP, short for Enterprise ResourcePlanning, and other business fields. It can provide much more applicable andreasonable management strategies for enterprises, and optimized resources allocationstrategies, making maximization use of resources and minimization of cost. What’smore, it can improve service quality and satisfaction of customers, increase profits forthe enterprises. Popularly, the association rule mining algorithms consider every itemto be mined to have the same value in default, neglecting different items may bringdifferent profit. However, what the enterprises consider most is profit. So these datamining techniques haven’t set out from the customers’ interest. On the other hand, thedata to be mined is enterprise-level huge, so it may not be stored in memory altogether.When this happens, there will be many times scanning database with I/O operations,affecting the performance and efficiency of the system.In this paper, it makes a further research, and introduces the logistics informationsystem, data mining technology and related knowledge and concepts first, thenrefering to achievements of other researchers in this field, it comes up with a weightedassociation rule mining algorithm based on partition technique---WPARAlgorithm.By integrating this algorithm into SQL Server2008, it can mine data set in SQLServer in different parameters. SQL Server Analysis Service can do ETL of the dataset. Meanwhile, it executes the association rule mining algorithm provided by SQLServer in the same condition, to verify the better performance of WPARAlgorithm.WPARAlgorithm divides the database into several parts, making sure that everypart can be stored in memory altogether. Then it operates only in one part every timein the memory, avoiding too many I/O operation and increasing the speed of thealgorithm. At the same time, it compresses and optimizes the linking and pruning step, and avoids pattern match step, to improve the performance of the algorithm. What’smore, it applies the profit-weighted method into the mining. In accordance with theenterprises’ interest, profit-weighted method can treat every item differently based onthe profit it can yield. And it can meet the customers’ need and be mor useful to them.