Research On Association Rules Mining Based On Cloud Model

Cloud theory is a theory dealing with uncertainty, including cloud model,virtual cloud, cloud operations, cloud transform, uncertainty reasoning, etc. Cloud model is a model of the uncertain transition between a linguistic term of a qualitative concept and its numerical representation. Cloud model is the basis of cloud theory, it integrates the fuzziness and randomness of concept representation and the uncertain transition between qualitatives and quantitatives. Converting quantitative attributes into Boolean attributes is the general way for mining quantitative association rules. Traditional method can not get the easy to understand knowledge because it can not reflect the actual data distribution or the partition is too sharp. A new method—cloud transform, which uses many concepts represented by cloud model to fit the real distribution of data—is introduced. This method can reflect the distribution of data in that domain while keeping the soft boundaries. Therefore, the discovered association rules are also easy to understand.In order to discover strong association rules, attribute values are generalized at higher concept levels, allowing overlapping between neighbor attribute values or linguistic terms. And this kind of soft partitioning can mimic human being's thinking, while making the discovered knowledge robust. Combining concept partition based on cloud models and Apriori algorithm, which can show the validity of cloud association rules. Association rules based on cloud models can mine association rules, whose confidence and support were greater than the user-defined minimum value.The main research work and achievements are as follows:1 Analyzing the mining association rules for the moment, especially the concept partition of quantitative attributes. The mining association rules of quantitative attributes transform the domain of quantitative attributes through the sector division method into the boolean attributes, then using boolean mines association rules. But it is obvious that how to divide the domain of quantitative attributes.2 The cloud generators are developed by Matlab, including normal cloud generator and backward cloud generator. The key technology of normal cloud generator is normal random data which is produced by central limit theorem. Then, it demonstrates the principle of producing random data. Normal cloud generator can translate qualitative language into quantitative data. Backward cloud generator can convert quantitative data into qualitative language at any time, so which is applied in mining.3 An algorithm of association rules based on cloud models is given. This method can well soften the domain partition boundary, so the cloud association rules can be easily understood. Cloud transform is used to fit the data distribution space probability density function by using cloud model ,in another words, atomic concept is transformed into the value of cloud model's digital feature .4 The domain of quantitative attribute is divided by using normal cloud model. According to association rules based on cloud models, predictions are made.