物质扩散和热传导过程物理过程已经在信息过滤领域发挥了重要作用:基于物质扩散过程的推荐算法可以提供很高的准确度，但在推荐列表多样性方面表现一般。而基于热传导过程的算法具有很高的推荐多样性，然而，准确性却表现不佳。我们认为，热传导算法之所以表现不佳，其原因在于给非流行的产品过多的权重。因此应该适当地给予度信息不太小的产品一些推荐权重，也提出了相应的基于偏向热传导的推荐算法。尽管操作简单，但是算法的准确度却可以和混合算法[PNAS 107 4511(2010)]达到相当的准确度，而推荐列表多样性比混合算法的结果还要好。

    进一步对推荐列表中的产品进行分析发现，算法之所以表现好是以为既能将流行产品放到推荐列表的顶端，也能适当地将冷门产品放在推荐列表的顶端。这与Facebook中用户的兴趣可以分为两大类（大众都喜欢的流行产品和自己独特喜好的冷门产品）的结果相吻合。

    Information filtering via biased heat conduction

    Heat conduction process has recently found its application in personalized recommendation [T.Zhou et al., PNAS 107, 4511 (2010)], which is of high diversity but low accuracy. By decreasing the temperatures of small-degree objects, we present an improved algorithm, called biased heat conduction (BHC), which could simultaneously enhance the accuracy and diversity. Extensive experimental analyses demonstrate that the accuracy on MovieLens, Netflix and Delicious datasets could be improved by 43.5%, 55.4% and 19.2% compared with the standard heat conduction algorithm, and the diversity is also increased or approximately unchanged. Further statistical analyses suggest that the present algorithm could simultaneously identify users' mainstream and special tastes, resulting in better performance than the standard heat conduction algorithm. This work provides a creditable way for highly efficient information filtering.