Abstract:

In order to improve the query quality of the traditional FTP search engines possessing low optimization performance for query results, a high-performance intelligent FTP search engine is designed based on the statistical analysis of FTP user query logs. In this engine, the double-byte inverted index is employed to build an inverted index table with every double bytes of the file name, the automatic classification of query results is used to establish a tree structure of query results based on the search topic, and the automatic error correction is adopted to construct a spelling mistake dictionary with the high-frequency search keywords in user query logs. Query results in a real system indicate that the proposed scheme greatly improves the query efficiency and quality of a FTP search engine.

Key words: File Transfer Protocol, search engine, inverted index, automatic classification, automatic error correction