Product detail page recommendation method based on multi-source infor-mation
The product recommendations on existing product detail pages usually recommend similar products based on the brand,store,category,and other information of the main prod-uct.Homogenization is severe,and there are difficulties in cold starting new products.Based on this,this article proposes a product detail page recommendation method based on multi-source information,which can moderately generalize based on the intention of the main product,ex-tract features from comment information,strengthen and extend users'immediate interests,not only alleviate fatigue during browsing,but also alleviate the problem of difficult cold start of new products.Through comparative experiments,the experimental results show that the pro-posed method has a high mAP value and a higher NDCG.It performs well in recommendation ranking and has a good recommendation effect.
Personalized recommendationsE-commerce platformRecommended methodsMulti-source information
NETL
NSTL
万方数据
