A. Demiriz (Turkey)
: Path Mining, Traversal Pattern Mining, Web Logs
A SImple yet very efficient PATH mining algorithm (ASI PATH) is introduced in this paper to analyze the click stream data. As in the case of the preceding sequence min ing algorithm (webSPADE), ASIPATH requires only one full scan of the data and several partial scans. In parallel al gorithm design, it is very common to parallelize a serial al gorithm, but ASIPATH is designed for multi-CPU environ ment in the first place without a preceding serial version. Due to the choice of the application environment i.e. Win dows, ASIPATH is not designed as a distributed algorithm but can easily be modified for a distributed environment. In contrast with many existing path mining algorithms, it is not based on a tree-like data structure and search method. By using very efficient join operations, the parallelization of the algorithm is simplified considerably.
Important Links:
Go Back
