基于图像和GM-PLSA模型的物品推荐方法  

Product Recommendation Based on Image Content and GM-PLSA Model

作　　者：朱韶平[1] 夏利民[2] 彭东亮[2] 

机构地区：[1]湖南财政经济学院,湖南长沙410205 [2]中南大学信息科学与工程学院,湖南长沙410075

出　　处：《系统工程》2013年第12期109-115,共7页Systems Engineering

基　　金：湖南省科技计划项目(2012FJ3021);国家自然科学基金资助项目(50808025);湖南省普通高等学校教学改革研究课题(湘教通【2012】401号544);湖南省重点学科建设项目

摘　　要：针对G-PLSA(Gaussian Probabilistic Latent Semantic Analysis)推荐模型中存在的新物品问题,提出了基于GM-PLSA(Gaussian Multinational Probabilistic Latent Semantic Analysis)模型的物品推荐方法;该方法将物品的图像内容特征用于物品推荐,有效解决了新物品问题和文本描述物品外观存在歧义性问题。同时,针对推荐结果的总体多样性问题,提出了ITRM(Improved Traditional Ranking Model)排序模型;该模型能够保证一定准确性的前提下有效提高推荐结果的总体多样性,且该模型具有参数化的优点。实验结果表明,本文提出的方法能有效提高物品推荐系统的推荐性能。In order to overcome the problem of new products existing in the G-PLSA （Gaussian Probabilistic Latent Semantic Analysis） model, a new recommendation method based on image content and GM-PLSA （Gaussian Multinational Probabilistic Latent Semantic Analysis） model is proposed ,which can solve the problem of new products and the problem of ambiguity existing in the way of presenting the product＇s appearance using text words effectively by combing the products their image features. A new re-ranking model called ITRM （Improved Traditional Ranking Model） is also proposed to improve the aggregate diversity for all users in the system. ITRM model can improve the aggregate diversity effectively with only a small amount of accuracy loss, and it is parameterized. The experimental results showed that the methods proposed in this paper can improve the recommendation quality of the product recommender system effectively.

关 键 词：物品推荐 图像内容 GM—PLSA模型 ITRM模型 总体多样性 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]