Study On Multi-dimension Of Temporal Association Rules Mining

With the rapid development of modern information technology, a great deal of data has been accumulated in many fields. People expect to discover the knowledge and rules existing in these data, which just brings the study of data mining and the development of its technology. As a comprehensive field of crossing multi-subject, data mining involves many subjects such as database, statistic, machine learning, high performance computing, pattern recognition, neural network and data visualization etc. Data classification and prediction are important mining technologies and have been used widely. The research of temporal data mining has become a hotspot. And temporal association rules discovery is an important field in data mining due to its wide applications. Currently, the study of temporal association rules mainly deal with the intra dimension associations and multidimensional association rules fewer take care of time. There exists some inadequacy in temporal association rule mining. Therefore, in this paper, we will study the problem of discovery of temporal association rules from the following aspects. ⑴The development of temporal association rules are discussed, then problems of them are presented. ⑵Based on the concepts and properties of time granularities, a model of multi-dimensional of temporal association rules are putted forward. It is applicable within the intra dimension associations and the inter dimension associations. ⑶Study the theories and algorithm of multi-dimensional of temporal association rules and the experimental results on the stock data are shown. ⑷Based on the theories of cyclic association rules, a new concept-approximately cyclic association rules is putted forward. And algorithm of this association rules is given. The experimental results on the stock data are shown in the end. In this paper, we obtain the following results. Firstly, we put forward the theories and algorithm of multi-dimensional and multiple granularities of temporal association rules. Secondly, we put forward the algorithm of life's interval length based on life's span. Thirdly, we put forward the concept, theories and algorithm of approximately cyclic association rules based on the theories of cyclic association rules...