基于SURP模型的物品推荐  

Product Recommendation Based on SURP Model

作　　者：谭论正[1] 夏利民[1] 彭东亮[1] 

机构地区：[1]中南大学信息科学与工程学院,湖南长沙410075

出　　处：《信息与控制》2014年第1期37-42,共6页Information and Control

基　　金：国家自然科学基金资助项目(50808025);高等学校博士学科点专项科研基金资助项目(20090162110057)

摘　　要：为了克服传统物品推荐技术中存在的局限,提出了一种基于SURP(supervised user rating profile)模型的物品推荐方法.利用词包(BOW)的方法,以图像特征来表示物品;在此基础上,采用监督学习方法来建立SURP模型,提高了对物品评分等级预测的准确性;通过引入用户兴趣因子,解决了用户对已购买物品的兴趣变化问题.在自建的物品数据集上,对此方法、URP(user rating profile)模型、G-PLSA(Gaussian probabilistic latent semantic analysis)模型和IBCF(item-based collaborative filtering)4种方法进行了对比实验.实验结果表明,该方法具有良好的物品推荐品质.To overcome the shortcomings of the traditional method in product recommendation, a new recommenda- tion method based on a supervised user rating profile （SURP） model is proposed. The products are represen- ted by their image features using the bag-of-word method. The SURP model is proposed using the supervised learning method, which can effectively improve the precision of the predicted ratings. The interest factor of the users is introduced to deal with the change in the interest of the users on the products bought previously. The method is tested on a product dataset. The experimental results show that the new method has a better recommendation capacity than the user rating profile, Gaussian probabilistic latent semantic analysis, and i- tem-based collaborative filtering methods.

关 键 词：URP(user RATING profile)模型 图像特征 物品推荐 用户兴趣 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]