Abstract:

The traditional recommendation algorithm has been widely used in various recommendation systems. However, there are still some problems in the recommendation process, such as inability to deal with unstructured data, difficulty in discovering data potential relationship, data sparsity and cold boot. The emergence of association rule technology alleviates the above problems and improves the efficiency of recommendation. The research on the high-quality combination between the special attributes of association rule technology and recommendation algorithm has become a hot topic. The application of association rule technology and different association rule categories of data in traditional recommendation algorithms are reviewed. The advantages and disadvantages of traditional algorithms in the process of recommendation are summarized. Finally, the research of association rule-based recommendation algorithm is summarized and the future development trend is predicted.

Key words: association rules, data mining, recommendation algorithm, collaborative filtering

摘要：

目前传统推荐算法已经被广泛应用于各类推荐系统，然而在推荐过程中仍然存在着无法处理非结构化数据、数据潜在关系发现困难、数据稀疏和冷启动等问题。关联规则技术的出现有效缓解了这些问题，推荐效率也因此得到提高。将关联规则技术的特殊属性与推荐算法进行高质量的结合成为推荐领域的研究热点。通过综述关联规则技术与数据的不同关联规则类别在传统推荐算法中的应用，对传统算法在推荐过程中的优缺点进行了归纳阐述。针对基于关联规则推荐算法的研究进行总结，并对其未来的发展趋势进行展望。

关键词: 关联规则, 数据挖掘, 推荐算法, 协同过滤