Social steams such as social network news feed have become major information sources for users. But the timeline sorting meth-od and the existing research and application are still not enough to solve the contradiction between"information overload" and the users"information hunger". This article first analyzed the existing recommendation models about social streams, then analyzed micro-blogs us-ers' needs and motivation, and discovered it can be divided into access to"social information","learning information" and"public infor-mation" three categories. According to micro-blogs users' motivation, this article constructed personalized recommendation models in on-line social streams based on ties strength, topic relevance and trust dimensions. Taking"Sina micro-blogs" as an example, this article fi-nally discussed social streams recommendation service.%微博客类社会化网络媒介中用户发布的社会化网络信息流，已经成为用户获取信息的主要信息源，但时间线排序法及现有研究与应用尚不足以解决“信息泛滥”与用户“信息饥渴”之间的矛盾。在分析现有微博客信息流推荐模式的基础上，将微博客用户动机分为获取社交类信息、学习类信息和公共信息三大类，构建了由主题相关性、关系强度和信任三个维度协同推荐的基于用户动机的微博客信息流推荐模型。以新浪微博“智能排序”为例，对基于用户动机的微博客信息流个性化推荐服务进行改进。
展开▼
