基于独特型人工免疫网络的并行推荐算法的研究

Research of parallel recommendation algorithm based on idiotypic artificial immune network

讨论了协同过滤最近邻居用户集缺乏多样性而导致推荐质量降低的问题,提出了并行免疫推荐算法(PINR),该算法能在保持邻居用户最大多样性的基础上进一步提高算法实时响应速度,实验结果证明了算法的可行性、正确性和并行计算的优势。

This paper analyzed the problem of recommendation quality degradation resulting from the lack of the nearest neighborhood users set＇s diversity in Collaborated Filtering Algorithm, and then proposed the Parallel Immune Network Recommendation algorithm （ PINR）. The algorithm achieves improvement in the real-time response speed while maintaining the maximum diversity of neighborhood users set. The preliminary experiment shows the feasibility, correctness and the advantage in parallel computing of this algorithm.