移动电子商务环境下个性化情景推荐模型研究

Research on Personalized Context Recommendation Model in Mobile E-commerce Environment

[目的/意义]文章用树形结构图对影响移动商务用户即时信息需求的情景要素进行分类,提出一种融入用户个性化情景要素的推荐模型。[方法/过程]在该模型中,通过用户当前情景和历史评分数据计算出对每个用户即时信息需求影响最大的K个情景要素,以此构造用户个性化情景,并将其作为输入条件改进已有的基于情景的多维信息推荐算法进而进行项目推荐。[结果/结论]通过Movie lens与MBook Crossing数据集对该模型中改进的算法和其他3种算法进行比较,实验结果表明:该模型的预测精度较高,可有效提高用户满意度,更能有效地解决移动商务环境下的个性化推荐问题。

[ Purpose/significance]This paper classifies the context elements which influence users＇ immediate information demand in mobile commence by tree structure chart and proposes a recommendation model of mobile commerce which is integrated into users＇ personalized context elements. [ Method/process]In this model, users＇ current context and historical data are used to calculate the top K contexts that could influence the immediate information demand most. On this basis, the paper constructs users＇ personalized context, and input the context as the condition to improve the existing context-based multi-dimensional information recommendation algorithm and program recommendations. [ Result/conclusion] By Movie lens and MBookCrossing, the algorithm proposed in the present paper is compared with other three algorithms. Results show that the prediction accuracy of the model proposed is higher, which can improve users＇ satisfaction effectively and solve the personalized recommendation problem in the mobile com- merce environment.