The Tag Navigation recommendation with adaptive learning method

Social Tags are widely used in web 2.0, and they bring the new chance and challenge to the recommender system, which is used to help users deal with information overload and provide personalized services. There are three respects of work done in this paper: firstly, the n-gram based Tag Navigation is presented to provide the assistant support for tag retrieval; secondly, the Average Mutual Information based tag similarity measure is detailed, furthermore this kind of semantic relation is applied to the retrieval intention expansion; thirdly, an approach of ranking based recommendation is presented, and the adaptive learning mechanism is explored. The experiments verify above methods, and result shows the complex features adopted in the recommendation bring improvement by 13.39%.