Irst Index Improve The Research And Application

One notable feature of full text index is the provision of management and rapid query to massive unstructured data. The space efficiency of creating index and the query speed after the completion of building index are two major hot spots in the field of research. In this paper, we compare some existing models of full text index. We also introduce a new data model of full-text database - Inter Relevent Successive Tree (in the text referred as IRST), and some progress on recent research.First, the query speed in the index has increased. The paper first introduces a Double Sorted Inter Relevent Successive Tree (DIRST), which was improved from the IRST model. And by comparing the performance of latest researched Backwards Search Algorithm of DIRST and existing Forwards Search Algorithm of DIRST, prove that Backwards Search Algorithm of DIRST is the most rapid query method on IRST.In addition, the article also introduced the application of IRST in frequent pattern mining. Through a model of Inter Relevent Successive Graph, the IRST model has been successfully used in frequent pattern mining. Comparing with the classic FP-Growth algorithm, IRST frequent pattern mining algorithm is generally much better than FP-Growth algorithm in most test cases, which proved that the IRST also has a very good prospect in frequent pattern mining.