The Research On Sequential Patterns Mining Based On Distributed Concept Lattices

As the result of Artificial Intelligence, Machine Learning, Databases and Statistics and other subjects, the aim of Data Mining and KDD is to find meaningful patterns from databases. So it is of great value. Sequential Pattern Mining is one branch of data Mining. With the quickly development of information technology, mining meaningful patterns effectively from large databases have become a challenge. An effective way to solve the problem is parallel or distributed technology. Duing to stability theory base and complete structure and parallel characteristic, concept lattice is an important tool to knowledge. In this dissertation, the model of distributed concept lattice and data mining research on which is studied.The main context is as follows:1. The research trends and the application of the data mining and KDD is introduced.2. The classical sequential patterns mining algorithms such as AprioriAll, GSP andPrefixSpan are discussed and compared. The Sequential Pattern Mining based on concept lattice are described.3. A new algorithm of sequential patterns based on Distributed Concept Lattice andSEA are proposed. All these are built on larger databases and SEA, to solve the problem of Sequential Patterns Mining for larger databases.4. Research on the certainty of sequential patterns mining. An algorithm of sequential patterns based on high certainty are proposed . The two different methods was studied, the one is sequential patterns mining before certainty used, the other one is sequential patterns mining before certainty used. The result is that the first one's efficiency was high.