User Interest Model at Multiple Dimensions and Multiple Granularities
Hui Chen, Song Deng · 2016
Finding user's interest and pushing related resources to user would be the best solution to aim the problem of "Information Overload" and "Lost in Internet".This paper designs a user interest model with multiple dimensions and multiple granularities, named UIM.In UIM, the interest concepts are kept by a hierarchical concept tree at different granularities, the history of interest are maintained with titled time window at multiple time granularities, and moreover current interest are discriminated from history ones by gradually fading the weights of interests.The results of simulations show that UIM is efficient to be maintained, can quickly extract users important interests in different time windows, and the correctness of finding user interest is prior to analogous methods.