Research On The Data Mining Model Based On Model Driven Architecture

With the rapid development of information technology,data acquisition and storage technologies have been widely used in people's daily life, a lot of data has accumulated in many areas. The data is being acquired as an exponential rate. However, only part of the data that has been transformed into knowledge, resulting in an embarrassment with data rich but poor knowledge. so data mining was put forward. Data mining mainly depend on the large databases or data warehouse to discover and extract hidden information or knowledge. which help analysts find the potential correlation between the data and found ignoring the elements, So it is very useful in predict information and decision making behavior.At present ,the modeling and design of DM depend on the specific structure of table and data warehouse of the business. The achievement of modeling and design of ARDM is hard to reuse, a new modeling and design of ARDM needs to re-analyze and re-design.The architecture of the MDA, meta modeling framework, as well as meta-modeling were researched in detail. On this basis, because of the association rules is so important in data mining research ,we designed the MDA based framework for mining association rules. Using UML extension mechanism designed association rule mining PIM, with the use of CWM we designed the association rules platform-dependent model of PSM.Established the source document which is the instance of PIM meta-model by analysising the sale fact of the Northeast special steel company. developed the rules based on the QVT transformation, so the source can be automaticly transformed into the instance of the PSM which is named target. Finally, in the EMF framework, by imported the target into EMF, we convert it into SQL code.As the model is based on internationally accepted standards, so it has a good normative. Meanwhile, it profits the re-use and the efficiency of system development.