Research On Pruning Techniques Of Mining Weighted Negative Sequential Patterns

Thanks to development of database technology and relevant technologies,various industries can access these technics to collect data,which results in a large number of digital resources storing the database.We urgently need to transform these digital resources into the useful knowledge,and let valuable knowledge can be utilized in the decision-making process of human-life.At present,the sequential patterns mining is a hotspot of data mining,which builds on the sufficient theoretical research.So that a variety of groundbreaking sequential patterns mining algorithms using in practical applications have been proposed.In the early stage of sequential patterns mining research,we did not consider the importance value of each item.If all the items to be regarded as the same important and it is not possible to ensure that which sequential pattern is the most interested of users.So,in this paper,a new weighting assignment method is proposed,which limits the range of weights reasonably.Sequential patters mining has some limitations,valuable information may both exist in occurred events and non-occurred ones.For this reason,negative sequential patterns mining was proposed,which requires more storage space and more running time.This paper summarizes the similarities and differences among classical positive and negative sequential patterns mining algorithms,and analyze the strategies and methods of pruning process through intensive experiments.In this paper,minimum support number and k-weighted support pruning strategies are involved in weighted negative sequential patterns mining algorithm WNGSP.All data set are coming from IBM data generator and UCI's official website.We compared the improved mining algorithm k-WNGSP and WNGSP for their performance test,the experimental results show that there is a difference between the running time and the number of the weighted negative sequential patterns obtained.Keeping all the other situations the same while doing experiments,the new algorithm can dig out much more weighted negative sequential patterns with less time.In summary,the k-WNGSP algorithm performs effectively and efficiently and gets the ideal experimental results on both three experiments.