电子商务平台大数据个性化推荐方法研究

Research on Personalized Recommendation Method of Big Data on E-commerce Platform

随着电子商务技术的快速发展,电子商务平台已成为企业推荐商品、用户实现商品消费的主要场所。无数企业将数以万计的商品在电商平台上呈现给用户,供用户选择,但这给用户选购商品带来了不便,同时也出现信息过载等问题。为解决该问题,目前电子商务平台大都应用了个性化推荐技术和算法,但在实际应用过程中,个性化推荐系统也存在稀疏矩阵、冷启动和用户差异等问题。为有效改善存在的问题,提升推荐质量,本文基于大数据和电子商务个性化推荐基本理论,构建了一种应用商品属性值和巴氏系数的优化协同过滤推荐算法,并对比分析了所构建的推荐算法与四种常见的推荐算法的推荐效果,验证了所构建推荐算法的有效性。

With the rapid development of E-commerce technology,E-commerce platform has become the main place for enterprises to recommend commodities and for users to realize commodity consumption.Countless enterprises present tens of thousands of commodities to users on E-commerce platforms for them to choose,but this brings inconvenience to users to choose commodities,and also causes problems such as information overload.To solve the problem,most E-commerce platforms have applied personalized recommendation technology and algorithm,but in the practical application process,personalized recommendation system also has sparse matrix,cold start,user differences and other problems.To effectively tackle the existing problems and improve the quality of recommendation,this article drew on large data and basic theory of E-commerce personalized recommendation,built an optimized collaborative filtering recommendation algorithm by applying commodity attribute values and pap coefficient optimization,compared and analyzed the recommendation effect of the built recommendation algorithm with four common recommendation algorithms,and verified the effectiveness of the built recommendation algorithm.