FairMatch: A Graph-based Approach for Improving Aggregate Diversity in Recommender Systems

Recommender systems are often biased toward popular items. In other words,\nfew items are frequently recommended while the majority of items do not get\nproportionate attention. That leads to low coverage of items in recommendation\nlists across users (i.e. low aggregate diversity) and unfair distribution of\nrecommended items. In this paper, we introduce FairMatch, a general graph-based\nalgorithm that works as a post-processing approach after recommendation\ngeneration for improving aggregate diversity. The algorithm iteratively finds\nitems that are rarely recommended yet are high-quality and add them to the\nusers' final recommendation lists. This is done by solving the maximum flow\nproblem on the recommendation bipartite graph. While we focus on aggregate\ndiversity and fair distribution of recommended items, the algorithm can be\nadapted to other recommendation scenarios using different underlying\ndefinitions of fairness. A comprehensive set of experiments on two datasets and\ncomparison with state-of-the-art baselines show that FairMatch, while\nsignificantly improving aggregate diversity, provides comparable recommendation\naccuracy.\n