2025春: R语言大作业(社交媒体平台用户数据分析)

社交媒体平台用户数据分析

摘要: 本文针对社交媒体平台用户与博主的互动行为展开分析, 数据源自 2025 年五一杯数学建模竞赛 C 题. 研究包含两部分内容:
其一, 基于 2021 年 7 月 11 日至 20 日的用户互动数据, 运用时间序列移动平均法, 对博主未来新增关注数进行预测, 筛选出预测新增关注数最多的 5 位博主, 并对模型的优势、局限性及优化方向进行探讨;
其二, 依据用户历史活跃率 (设定阈值为 50%) , 预测指定用户在 2021 年 7 月 21 日的在线状态, 若在线, 进一步通过互动次数结合时间新近性, 预测该用户可能产生最高互动数的 3 名博主, 同时分析模型的优缺点及优化路径, 如引入时间衰减模型、行为类型加权、结合用户画像等. 研究为平台优化推荐算法、提升用户互动提供了数据支持与方法参考.

关键词: 社交媒体分析; 用户行为预测; 移动平均法; 互动推荐; 活跃率

Analysis of User Data on Social Media Platforms

Abstract: This paper analyzes the interaction behavior between users and bloggers on social media platforms, with data from Problem C of the 2025 Wuyi Cup Mathematical Modeling Competition. The research consists of two parts:
Firstly, based on the user interaction data from July 11th to 20th, 2021, the time-series moving average method is used to predict the future new follow-up numbers of bloggers. The top 5 bloggers with the highest predicted new follow-ups are identified, and the advantages, limitations, and optimization directions of the model are discussed.
Secondly, according to the user's historical activity rate (with a threshold set at 50%), the online status of specified users on July 21st, 2021 is predicted. If online, the top 3 bloggers with the highest possible interaction numbers for each user are further predicted by combining interaction frequency with temporal recency. The paper also analyzes the advantages, disadvantages, and optimization paths of the model, such as introducing time-decay models, behavior-type weighting, and integrating user portraits. This study provides data support and methodological references for platforms to optimize recommendation algorithms and enhance user interaction.

Keywords: Social media analysis; User behavior prediction; Moving average method; Interaction recommendation; Activity rate

以下是文章:
社交媒体平台用户数据分析