An Evaluation Model for Degree of Topic Relatedness Based on Dynamic Matching

An evaluation model for degree of topic relatedness based on dynamic matching is proposed in this paper, which is based on web page segmentation and primary content block detection and is aimed at large scale web page collections. The evaluation model can suit automatically with inputted data type and is able to resolve partly the contradiction between time-consuming and accuracy rate. In addition, term weighting formula and nonlinear computing function which are the basis of the proposed evaluation model are improved in this paper after a lot of studies and experiments. The experimental results show that the evaluation model proposed in this paper performed better than present approaches.