Popularization mode of government-dominated rural residential demonstration projects and its features were introduced.It was proposed that such a popularization mode still remained at the primary development stage and...Popularization mode of government-dominated rural residential demonstration projects and its features were introduced.It was proposed that such a popularization mode still remained at the primary development stage and was still suffering from many problems such as guiding role of government not fully played,lack of assessment standards,unitary funding channels,ineffective popularization effects and unpractical management regulations.In view of these problems,popularization strategies for rural residential demonstration projects were given from the perspective of establishing authorized coordination and supervision agencies,improving incentive policies of demonstration projects and also regulation system for these projects,expanding funding channels,and building information platforms.The study was to provide practical references for the future development of rural residential demonstration projects.展开更多
数据稀疏问题普遍存在于协同过滤系统,仅考虑共同评分项目上局部上下文信息的相似度度量方法已不具备较高可靠性。为解决上述问题,提出一种融合多语义信任度和全局信息的混合推荐算法(multi semantic trust and global knowledge,MSTGK...数据稀疏问题普遍存在于协同过滤系统,仅考虑共同评分项目上局部上下文信息的相似度度量方法已不具备较高可靠性。为解决上述问题,提出一种融合多语义信任度和全局信息的混合推荐算法(multi semantic trust and global knowledge,MSTGK)。引入加权异构信息网络(weighted heterogeneous information network,WHIN),通过加权元路径处理评分数据、社交关系、用户标签和项目属性对用户信任的影响,挖掘不同语义的信任信息以缓解数据稀疏性问题;考虑项目流行度和用户偏好程度两个全局要素对用户相似度的影响,将其作为权重因子改进了JMSD相似测度,旨在提高相似度计算精度;融合用户的多语义信任度和全局相似度进行综合推荐。在DoubanMovie和Yelp两个真实数据集上的实验结果表明,所提算法缓解了数据稀疏问题,相比于其他基线方法,预测准确率分别提高了2.01个百分点和2.45个百分点。展开更多
基金Supported by Key Project of National Scientific and Technological Support Program(2008BAJ08B16)~~
文摘Popularization mode of government-dominated rural residential demonstration projects and its features were introduced.It was proposed that such a popularization mode still remained at the primary development stage and was still suffering from many problems such as guiding role of government not fully played,lack of assessment standards,unitary funding channels,ineffective popularization effects and unpractical management regulations.In view of these problems,popularization strategies for rural residential demonstration projects were given from the perspective of establishing authorized coordination and supervision agencies,improving incentive policies of demonstration projects and also regulation system for these projects,expanding funding channels,and building information platforms.The study was to provide practical references for the future development of rural residential demonstration projects.
文摘数据稀疏问题普遍存在于协同过滤系统,仅考虑共同评分项目上局部上下文信息的相似度度量方法已不具备较高可靠性。为解决上述问题,提出一种融合多语义信任度和全局信息的混合推荐算法(multi semantic trust and global knowledge,MSTGK)。引入加权异构信息网络(weighted heterogeneous information network,WHIN),通过加权元路径处理评分数据、社交关系、用户标签和项目属性对用户信任的影响,挖掘不同语义的信任信息以缓解数据稀疏性问题;考虑项目流行度和用户偏好程度两个全局要素对用户相似度的影响,将其作为权重因子改进了JMSD相似测度,旨在提高相似度计算精度;融合用户的多语义信任度和全局相似度进行综合推荐。在DoubanMovie和Yelp两个真实数据集上的实验结果表明,所提算法缓解了数据稀疏问题,相比于其他基线方法,预测准确率分别提高了2.01个百分点和2.45个百分点。
