基于标签的商品推荐模型研究

发布时间：2018-09-17 10:04

【摘要】：【目的】构建社会化电子商务环境下基于标签的个性化商品推荐模型。【方法】综合考虑用户使用标签的频率和时间因素计算用户的兴趣偏好;基于标签层次特征和电子商务网站中关于商品特征的检索条件,构建某一主题商务社区中商品本体;利用本体规范化用户标签语义,并对商品进行分类;寻找含有用户偏好的类簇,计算该类簇中商品与用户偏好商品的相似度,将用户未标注过的商品与用户偏好相似度高的商品推荐给用户。【结果】从翻东西网站上随机选取200个活跃用户关于热门商品的标注信息进行分析,验证该模型的有效性。【局限】在计算用户兴趣偏好时,只考虑用户使用标签的频率和时间因素,未考虑其他因素。【结论】该模型相对于利用标签进行协同过滤推荐方法具有较优的效果,计算时间和空间复杂度更小。
[Abstract]:[objective] to construct a model of personalized commodity recommendation based on label in the environment of social electronic commerce. [methods] considering the frequency and time of users' use of labels, we can calculate users' preferences of interest. Based on the hierarchical features of labels and the retrieval conditions of commodity features in e-commerce websites, the commodity ontology in a subject business community is constructed, and the user label semantics are standardized by ontology, and the goods are classified. In order to find the cluster with user preference, the similarity between the products in the cluster and the user preference is calculated. The untagged items were recommended to users with high similarity to users' preferences. [results] 200 active users were randomly selected to analyze the tagged information about hot items on the website. Verify the validity of the model. [limitations] in calculating user preferences, only the frequency and time of the user's use of the label is considered, No other factors are considered. [conclusion] this model has a better effect than the collaborative filtering and recommendation method using tags, and the computational time and space complexity are smaller.
【作者单位】：东华理工大学经济与管理学院;武汉大学信息管理学院;
【基金】：国家自然科学基金项目“社会化媒体集成检索与语义分析方法研究”(项目编号:71273194) 抚州市社科规划项目“基于KANO模型的抚州旅游市场需求分析”(项目编号:15sk23) 东华理工大学地质资源经济与管理研究中心开放基金项目“基于KANO模型的旅游用户潜在需求挖掘研究”(项目编号:14GL05)的研究成果之一
【分类号】：TP391.3

【相似文献】