协同过滤算法在微信推荐小程序的应用  

Simulation Modeling and Analysis of Ocean Transportation System under Sea Ice Condition

作　　者：刘彦会[1] LIU Yanhui(Minbei Vocational and Technical College,Nanping,Fujian 353000)

机构地区：[1]闽北职业技术学院,福建南平353000

出　　处：《武夷学院学报》2024年第6期51-57,共7页Journal of Wuyi University

摘　　要：为了使客户能够方便、快速从大量数据中获取有效信息,本文对微信小程序采用Mvc的开发模式,并以Node.js技术进行设计,其总体架构主要包括交互层、数据访问层、控制层和数据库层。采用基于用户的协同过滤推荐算法和基于特征的协同过滤推荐算法结合的方式进行商品的推荐。对于基于用户的算法,采用IG特征选择算法进行商品特征的选取,再采用改进的Pearson相关系数进行相似性计算,获取推荐商品。对于基于特征的算法,采用改进的余弦相似性进行用户相似度的计算,根据相似性用户推荐商品。将推荐商品按照比例结合,最终进行商品的推荐。为了验证该微信小程序的性能,对其进行微信推荐小程序运行测试和商品推荐测试。试验结果表明微信小程序的各移动端均可正常运行,各项功能可进行操作,且向用户推荐的有效信息符合设计要求。The traditional use of mobile phone for information query,restaurant ordering,commodity purchase need mobile phone download APP or iPad for the use of related functions,cumbersome process,occupy too much memory of the phone.In order to enable customers to easily and quickly obtain effective information from a large amount of data,the Mvc development mode and Node.js technology were adopted to design the wechat small program.The architecture of the wechat small program was mainly included interaction layer,data access layer,control layer and database layer.The product recommendation was carried out by combining user-based collaborative filtering recommendation algorithm and feature-based collaborative filtering recommendation algorithm.For the user-based algorithm,the IG feature selection algorithm was used to select the commodity features,and then the improved Pearson correlation coefficient was used for similarity calculation to obtain the recommended commodities.For the feature-based algorithm,the improved cosine similarity was used to calculate the user similarity,and products were recommended according to the similarity.Combine the recommended goods according the proportion,and finally carry out the product recommendation.In order to verify the performance of the wechat small program,the wechat recommendation small program operation test and commodity recommendation test were carried out.The test results show that all mobile terminals of wechat small program could operate normally,all function could be operated,and the effective information recommended to users meets the design requirements.

关 键 词：微信小程序 协同过滤算法 信息增益(IG)特征 商品特征 Pearson相关系数 

分 类 号：TP391.3[自动化与计算机技术—计算机应用技术]