基于物品相似度的异类商品推荐算法研究
Research on Heterogeneous Product Recommendation Algorithm Based on Item Similarity

现有的推荐算法大多是基于同类商品进行推荐，容易形成“信息茧房”。为了解决同类商品推荐的局限性，将推荐算法引申到不同类别的商品推荐之中，本文提出了一种基于物品相似度的异类商品推荐算法。在应用了物品相似度的基础上，提出交叉相关推荐理论，解决了目标商品与推荐商品集异构推荐的问题。最后，本文从天池淘宝穿衣搭配数据集中提取商品数据，通过对所提出的算法设计程序语言，应用到数据集中进行分析。根据所得到的实验结果，该算法得到的推荐成功率较高，推荐效果较好。
Most of the existing recommendation algorithms are based on the recommendation of similar products, which can easily lead to “information cocoon rooms”. In order to solve the limitations of similar product recommendation, the recommendation algorithm is extended to different catego-ries of product recommendation. This paper proposes a heterogeneous product recommendation algorithm based on item similarity. Based on the application of item similarity, a cross-correlation recommendation theory is proposed to solve the problem of heterogeneous recommendation of target products and recommended product sets. Finally, this article extracts the product data from the Tianchi Taobao clothing matching data set, and applies the proposed algorithm to the pro-gramming language to analyze the data set. According to the obtained experimental results, the al-gorithm has a high recommendation success rate and a good recommendation effect.