摘要
为解决微博信息过载问题,信息过滤、推荐和搜索等技术被应用于博客研究中。但已有推荐算法多数基于单一维度进行推荐,导致推荐场合存在局限性。为此,构建一种综合信任模型、社会关系和语义相似度的综合推荐模型。应用神经网络来学习和捕获博主的偏好模式,并且将其用于预测最终的推荐得分。在真实的数据集上进行实验,结果表明该推荐机制具有高效性与可行性。
Approaches of messages filtering, recommendation and searching are proposed to solve the problem of blog information overload. Most existing recommendation algorithms are based on single dimension recommendation,which leads to limitations in recommendation situations. To solve this problem,a comprehensive recommendation model including integrated trust model, social relation analysis and sematic similarity is built. Neural network is applied to learn and capture bloggers＇ preference pattern, and it is used to predict the final recommendation score. Experimental results on real data sets show that the recommendation mechanism is efficient and feasible.
作者
林杰
刘波
LIN Jie;LIU Bo(School of Public Administration,South China University of Technology,Guangzhou 510640,China;School of Computer Science,South China Normal University,Guangzhou 510640,China)
出处
《计算机工程》
CAS
CSCD
北大核心
2018年第7期271-278,共8页
Computer Engineering
基金
国家自然科学基金青年项目(513060611)
广东省哲学社会科学规划项目(GD16YGL02)
关键词
信任度
社会网络
语义
博客
推荐
degree of trust
social networks
semantics
blog
recommendation
