摘要
在电子商务的商品推荐系统中,为了给用户提供个性化的商品推荐,不少研究者提出了各自的推荐方法.由于方法中考虑的影响因素小等原因使得一些问题仍未得到解决,如:精准度不高、时间复杂度较高等.针对以上问题,在LDA主题挖掘模型的基础上提出了一种新的适用于商品(SP)推荐的数据挖掘模型:SP_LDA模型.通过该模型进行推导,得到商品概率的计算公式.通过对用户的历史购买数据和浏览数据进行分析,以计算的方式求解商品被推荐的概率,最终得到用户潜在感兴趣的商品.实验表明本模型能够高效地对商品进行挖掘,合理地向用户推荐感兴趣的商品.
In commodity recommendation system of e-commerce,in order to provide users with personalized commodity recommendations,many researchers have proposed their own approches. Some problems remain unresolved due to fewinfluential factors and etc.were considered,such as： lowprecision and high time complexity. To solve the these problems,a newrecommendation model which is based on the LDA model is proposed,it is a data mining model which applies to recommending commodities（ SP） ： SP_LDA model.Through the derivation of the model,we get the calculation formula of the probability of goods. Users＇ historical purchasing data and browsing data are analysized in this model to calculate the probability of the recommended commodities,then we finally get commodities which the user is interested in. Experiments showthat the SP_LDA model can mine commodities effectively and recommend commodities to users reasonably.
出处
《小型微型计算机系统》
CSCD
北大核心
2016年第3期454-458,共5页
Journal of Chinese Computer Systems
基金
国家自然科学基金项目(61379057)资助
