Research And Implementation Of OLAP And Data Mining Algorithms In Retail Industry

Data mining(including OLAP) is an subject which associates with practice closely. Faced with more and more data storing every year in almost all kinds of domains, such as retail, finance, science and medical, there is an urgent need for data mining to use its kernel thought-have an insight to data and make discovery-to real application.It's a pity that although there are many papers and articles focused on data mining published every year, most of them deal with data mining concept and abstract algorithm theory, it is hardly to see their real implementation and application, In this context, when I was in my graduate exercitation in a company in Beijing, which focus on developing supermarket software, I joined and completed an OLAP(Online Analytical Processing) project, merchandise analysis and sale report system, which based on Microsoft Analysis Service and Microsoft Sql Server. I also design and implement three important algorithms: merchandise association rule algorithm based on multi-level merchandise category, supermarket member customer shopping frequent sequence generating algorithm, customer classification(decision tree) algorithm which based on information entropy and conditional probability tree, and they all achieve expected result. The first and the third achieve good performance and scalability, the second is effective when used in small and medium datasets.