Abstract: This paper describes the goal of relevance algorithm and relevance calculation connotation in focused crawler. Then, according to the evolutionary point of view of information processing, it systematically analyzes the current relevance calculation method of focused crawler in three levels: character layer, language layer, semantic layer, and compares the advantages/disadvantages among algorithms from different levels. Finally, it summarizes the current research results and indicates the direction in future works.

Key words: relevance, algorithm, focused crawler, concept