所属单位：计算机科学与工程学院

发表刊物：The 4thInternational Conference on Computer Engineering and Network

项目来源：省、市、自治区科技项目

关键字：TF-IDF； VSM ； Feature extraction ； Weight calculation； Topic filtering

摘要：Abstract Feature weighting algorithm is a key step in feature extraction. In this paper，according to two classifications situation of web page theme, we improved the traditional feature item weighting method of TF-IDF. On the basis of this, We give a new linear feature weighting algorithm based on the feature’s position information and the information it contained. Experimental results show that this algorithm has higher the theme web page recall rate and precision rate than the traditional method TF-IDF in the theme web page filtering.

合写作者：田世元,王国春,代宽,赵辉

第一作者：赵辉

论文类型：论文集

页面范围：2

是否译文：否

发表时间：2014-07-19