Keyword weight adjusting schema based on Domain Repository

Webpage keyword is widely used in personalized search and recommendations. The accuracy of keyword is significant to the quality of search and recommendation result. Current keyword extraction methods´ accuracy is not high. In order to make up for the shortage of present technology, a new keyword weight adjusting schema based on Domain Repository has been proposed in this paper. Three steps: 1) Extract original keywords by word frequency statistics. 2) Match original keywords with concepts in Domain Repository in this page´s field. 3) Increase the weight of words more related to this field then abstract its meaning to build the weighted keywords. Extensive experiments show that this method has better resource description accuracy than traditional statistical keyword extracting method.