Algorithm For Mining Weighted Negative Sequential Patterns

The traditional sequential pattern mining is to find frequent sequences from a sequence database as a model of knowledge discovery process, which is an important data mining research, there are practical applications in many fields. In recent years, domestic and endless sequence pattern mining research provides a number of solutions for business, medical field analysis of decision problems. However, some important information may be hidden in the sequence of the missing element. Therefore, the negative sequential pattern mining technology has begun to enter the field of vision of scholars.Currently, the research of negative sequential pattern mining techniques in the fledgling stage, the relevant theoretical results are lackluster. And look at the negative sequential patterns mining algorithms have been proposed, most of them exist blindness, words simply tap the negative sequence mode, and not to dig out the negative sequence weighting pattern, which do not reflects the importance of the sequence of the project, and the negative sequence mode will appear in the case of an existing algorithm of combinatorial explosion. Aiming at two problems were discussed above, this paper introduce the negative sequence pattern mining in setting reasonable weight value and introducing a control mechanism to set the value of the right to post negative sequence weighting pattern, based on the above this paper proposes a the new species weighted negative sequence pattern mining algorithm with comparison of unweighted, it reflects the importance of the concept of weighting and real data sets operating environment experimentally verify the effectiveness and feasibility of the algorithm.In addition, this paper also introduces the concepts and methods of data mining as well as positive and negative sequential pattern mining, and through the experimental comparison draw some useful conclusions.