摘要
推荐系统已经在日常生活中扮演着举足轻重的角色,单一的推荐系统往往会存在冷启动、数据稀疏等问题,该文将各推荐服务的结果通过动态权重的方式加以调整并混合,避免单一算法带来的问题,提升个性化推荐效果。将大数据技术和推荐算法结合,设计并实现基于大数据的商品混合推荐系统。最后使用Amazon的数据集进行系统测试,该文提出的动态权重混合方式比传统线性混合拥有更好的性能。
Recommendation systems have played an important role in daily life.A single recommendation system often has problems such as cold start and data sparseness.This article adjusts and mixes the results of each recommendation service through dynamic weights to avoid a single algorithm.To improve the effect of personalized recommendation.Combine big data technology and recommendation algorithm to design and implement a product hybrid recommendation system based on big data.Finally,use Amazon’s data set for system testing.The dynamic weight mixing method proposed in this article has better performance than traditional linear mixing.
作者
周双杰
高凤玲
孙知信
ZHOU Shuangjie;GAO Fengling;SUN Zhixin(School of Modern Posts,Nanjing University of Posts and Telecommunications,Nanjing,Jiangsu Province,210000 China;China Electronics System Engineering NO.2 Construction Co.,Ltd.,Wuxi,Jiangsu Province,214135 China)
出处
《科技资讯》
2021年第10期32-34,共3页
Science & Technology Information
基金
江苏省研究生科研创新计划(项目编号:SJCX19_0244)。
关键词
推荐系统
混合推荐
动态权重
个性化推荐
Recommendation system
Hybrid recommendation
Dynamic weight
Personalized recommendation