中文摘要：

针对个性化知识服务不够准确和完整的问题，通过引入相关反馈技术构建个性化知识服务模型，并进一步改进Rocchio反馈算法，提出一种根据用户显性和隐性综合反馈度来动态更新用户模型库的相关反馈机制，以向用户提供更加准确、完整的个性化知识服务。实验证明引入这种反馈机制能够提高个性化知识服务系统的准确性和完整性，是可行和有效的。

英文摘要：

In order to provide users with more accurate and complete personalized knowledge services, this paper builds a personal- ized knowledge service model by introducing the relevance feedback technology and proposes a kind of relevance feedback mechanism that can dynamically updates the user model library by a comprehensive feedback degree consisting of explicit feedback degree and im- plicit feedback degree. The paper also makes a further improvement based on the Rocchio feedback algorithm. Experiments prove that such a feedback mechanism is feasible and effective.