Method for digging frequency mode in the lately time window of affair data flow

Abstract

The present inventionprovides a method of mining the frequent pattern on the transaction data stream of the latest time window. The frequent pattern adopting the prefix pattern reusing manner compresses the frequent pattern on the storing data stream in the latest time window; when the stream data arrives, the contained pattern information increment is updated to the frequent pattern tree; at thesame time, a time attenuation model is used for undertaking the attenuation to the pattern supporting quantity contained in the history transaction data, and hereby, the newly produced transaction datapattern is separated from the history transaction data pattern, highlighting the importance of the new transaction pattern; when a user submits the pattern for searching, the method of the present in vention can response the searching immediately and output the latest frequent pattern on the data stream. Compared with the other methods, the method has the advantages of quick response speed and fine pattern mining time and so on, ensuring the covering rate or fineness of the pattern mining at the same time.

Description

technical field [0001] The invention belongs to temporal data mining technology, in particular to a method and system for mining frequent patterns in the latest time window on transaction data streams. Background technique [0002] A data stream is a class of unbounded data sequences composed of data elements arriving at high speed. In recent years, data streams have been widely used in various applications, such as: network traffic monitoring, financial data management, sensor network data management, Web log analysis, mobile object data management, communication data analysis, and transaction log management in large retail businesses. In such applications, it is of great significance to discover frequent patterns in transactional data streams. For example, in network traffic monitoring, frequent packets corresponding to abnormal traffic may mean that there is a network attack; in a large number of retail sales records, Frequent items and their combinations always correspo...

AI Technical Summary

Problems solved by technology

Nevertheless, this method still has the following defects: (a) the dictionary tree in this method cannot effectively utilize the storage space, and the space utilization rate is low

(d) This method needs to traverse the tree nodes on all paths on the dictionary tree when pruning the dictionary tree, and the efficiency is relatively low

Images