社会化标签系统中个性化的用户建模方法

Method for personalized user profiling in social tagging systems

针对社会化标签系统中现有用户兴趣模型建立的缺陷,即:使用一些零散标签的集合来表示用户兴趣,而忽略标签的联合使用现象。提出一种将共现技术引入自然法的用户建模方法,该方法以自然法为基础,向用户模型中添加适量的标签对,较好地体现了标签之间的联系,又同时考虑了体现用户兴趣的标签自身的权重。在PKDD2009数据集上测试实验结果表明,该模型较之已提出的自然法和共现法,取得了更高的准确率和召回率。

As for the disadvantages of available user interests modeling methods in social tagging systems： always use a set of scattered tags to represent users＇ interests but ignore the combination use of tags,a new user modeling method was proposed,which introduced co-occurrence techniques into the naive approach.The proposed method added some tag pairs on the basis of naive approach to user model,which reflects tags relationship and considers tags weight.The experimental results on dataset PKDD2009 show that the new model can achieve a higher precision and recall rate,which outperforms the nave method and co-occurrence method.