In the last decade, the significant growth of the volume of analysis data has set the high level of importance of data mining field. This field contains a vast amount of different methods and techniques for knowledge extraction. One of the highly-demanded areas of this field is sequential pattern mining (SPM), which includes many methods for detection of frequent sequential patterns in different types of input ordered data sets. The goal of this work is to compare the efficiency of several types of SPM algorithms, and to identify the most applicable algorithm to deal with data from physical experiments used in scientific analysis tasks (e.g., analysis data from the ATLAS experiment at the Large Hadron Collider, CERN, Switzerland), and to extract association rules from experimental data samples. This paper presents the analysis of 3 types of SPM algorithms - horizontal and vertical, as well as pattern-growth, with the emphasis on algorithms? performance. There were prepared corresponding test data sets which are specific and typical for analysis tasks in the ATLAS experiment.
展开▼
