电子商务中基于潜在类回归模型的农产品个性化推荐方案

发布时间：2018-02-28 04:38

本文关键词： 电子商务农产品个性化推荐潜在类回归模型组群偏好　出处：《江苏农业科学》2017年12期 　论文类型：期刊论文

【摘要】：针对现有电子商务中农产品个性化推荐方案精度较低的问题,提出一种基于潜在类回归模型(latent-class regression model,简称LCRM)和组群偏好的个性化推荐方案。首先,收集农产品的评价信息,进行预处理,提取出每个评价者的特征-意见值对。然后,利用LCRM根据整体与特征评价,将具有相同爱好的评价者进行分组,构建组群偏好,并计算单个评价者的偏好。最后,通过计算用户与组群偏好的相似度来定位组群,通过计算用户与该组群中评价者偏好的相似度来定位农产品,最终列出推荐表。结果表明,该方案能够准确为用户推荐所需的农产品,推荐列表中农产品的命中率达到了83%,同时具有较低的计算复杂度。
[Abstract]:In order to solve the problem of low precision of individual recommendation scheme for agricultural products in electronic commerce, a personalized recommendation scheme based on latent class regression model (LCRM) and group preference is proposed. Firstly, the evaluation information of agricultural products is collected. After preprocessing, each evaluator's character-opinion value pair is extracted. Then, using LCRM to evaluate the whole and the characteristics, the reviewers with the same interests are grouped, the group preferences are constructed, and the preferences of the individual evaluators are calculated. Finally, By calculating the similarity between user and group preference to locate the group, and by calculating the similarity between user and evaluator in the group to locate agricultural products, the recommendation table is finally listed. The results show that, The scheme can accurately recommend the desired agricultural products to the users. The hit ratio of the agricultural products in the recommended list is 83% and the computational complexity is low.
【作者单位】：四川工程职业技术学院经济管理系;电子科技大学计算机科学与工程学院;
【基金】：四川省科技支撑计划(编号:2014RZ0048)
【分类号】：TP391.3

【参考文献】