CAJ | 학술논문

针对传统基于内容过滤的服务资源推荐技术中，用户对不同属性偏好权重的计算依赖于用户Web使用行为，不能适用于用户行为数据缺乏的场合，以及用户对属性值偏好计算缺乏用户评分反馈考虑等问题，提出了一种基于内容过滤的服务资源推荐新技术，该技术依据用户消费业务数据和评分反馈，通过计算服务资源属性的信息熵给属性权重赋值，并根据属性值变化频率对属性权重进行调整，得到用户对属性的偏好权重；根据属性值概率分布和用户评分，引入遗忘因子计算用户对属性值的偏好。实验结果表明，该技术建立的用户偏好模型和依据该推荐的服务资源结果更为准确。
침대전통기우내용과려적복무자원추천기술중，용호대불동속성편호권중적계산의뢰우용호Web사용행위，불능괄용우용호행위수거결핍적장합，이급용호대속성치편호계산결핍용호평분반궤고필등문제，제출료일충기우내용과려적복무자원추천신기술，해기술의거용호소비업무수거화평분반궤，통과계산복무자원속성적신식적급속성권중부치，병근거속성치변화빈솔대속성권중진행조정，득도용호대속성적편호권중；근거속성치개솔분포화용호평분，인입유망인자계산용호대속성치적편호。실험결과표명，해기술건립적용호편호모형화의거해추천적복무자원결과경위준학。
For the traditional recommendation technology of service resources based on content filtering , the calculation of the user's preference weights for different attribute depends on the user's past behavior in the Web .However , it is not appropriate for the situation lacking the user's past behavior data and the user's rating feedback when calculating the user's preferential attribute value.A new recommendation technology of service resources was proposed based on content filtering .It can obtain the user's preferential attribute weight , through the business data of the user's consumption and the user's rating feedback , assign attribute weights based on the information entropy of the service resource attributes , and adjust the attribute weights according to the change frequency of attribute values .The preference to the attribute value was calculated based on the probability distribution to the attribute values, the user's ratings, and the forgetting factor.The results show that the user's preference model and the recom-mended service resources according to the new recommendation technology are more accurate .