摘要
建模算法属于个性化搜索引擎的范畴,而空间向量建模是表示页面特征和用户兴趣的方法之一。建模的效率的好坏直接影响到个性化搜索的准确性。文章在基于统计学的基础上对特征词频率计算的作出了适当的调整,通过引入非线性函数使得特征词的权重非线性增加,更加反应了词语的频率的真实性。建模算法的改进更加考虑了特征词的所在的位置,提高用户兴趣向量的准确性,改进了计算页面特征向量和用户兴趣向量的相关性,从而提高了个性化兴趣度的准确性,使个性化搜化更加符合用户的兴趣。
Modeling algorithms belong to the scope of personalized search engine, while the space vector modeling is the ways of signifying page features and user interest. Efficiency of modeling directly impact on the accuracy of the personalized search. In this paper, based on statistics it mades appropriate adjustments of calculating of the frequency feature word, by the introduction of non-linear function making the weiht of feature word non-linear increase, more reaction to the authenticity of the feature word. Modeling algorithm improvement is more taken into account the location of the feature world and im- prove the accuracy of user interest vectors to improve relevance of page feature vector and the user interest vectors, thus im- proving the accuracy of the degree of personal interest, so that individual search more will be in line with user interest.
出处
《计算机与数字工程》
2010年第2期15-17,135,共4页
Computer & Digital Engineering
