摘要
论文采集腾讯微博数据,提出"博文魅力指数"的概念,并运用Spearman和Pearson相关系数分别对听众数与收录博主人数、博文魅力指数与收录博主人数两对变量进行了相关分析,最后选择博文魅力指数,博主收听人数两个变量使用K-Means聚类算法对微博用户进行了聚类分析。研究结果表明:博文魅力指数与收录博主人数两变量中度正相关;聚类将微博用户分为信息获取型、草根名人型和普通社交型三类。电子商务服务商可以通过算法优化,根据博文魅力指数和详细的聚类结果更有针对性的进行页面和应用程序推荐,创造商业价值。
Based on the data collected of Tencent Micro-blog, this paper proposes the concept of "the content' s charm index of micro- blog", then uses Pearson mad Spearman correlation coefficient to aimlyze the relatiemship between the number of listeners and the number of users, and the relationship between "the content' s charm index of micro-blog" and the number of users. At last, the paper uses K- Mealis clustering algorithm to analyze the characteristics of users. The conclusion is that the content' s charm index of micro-blog has mod- erate positive correlation with the number of users. K-Means clusteimg classifies the users into three types such as information-obtaining type, grassroots celebrities and ordinary social networking type. Thus, by optimizing the algorithm and using the content' s charm index of micro-blog and the result of clustering, service providers can reduce unnecessary recommendations of pages and application to meet the needs of users so as to acquire commercial value.
出处
《情报杂志》
CSSCI
北大核心
2012年第3期84-87,共4页
Journal of Intelligence