论文部分内容阅读
当前档案检索利用领域遇到的困难多可规约为三类问题:兴趣利用集问题、事务共同利用集问题、档案异名检索问题。这些问题的实质是档案数据难以智能化地到达目标用户。因此,引入个性化推荐的概念和技术,建立档案领域自己的数据推荐引擎是必要的。对于建设该引擎所面临的用户特点、数据特点和现实建设条件而言,协同过滤技术是十分适用的,但又不可对其机械照搬套用,应根据档案应用领域的实际特点进行变通。文章分别面向兴趣利用集问题、事务共同利用集问题、档案异名检索问题的具体特点,设计出各自的协同过滤矩阵和推荐实现过程。
The current difficulties encountered in the field of file retrieval and use of multi-protocol can be conventionally classified into three types of questions: the use of interest set, the common use of issues set, file retrieval. The essence of these problems is that it is difficult for intelligent data files to reach the target users. Therefore, it is necessary to introduce the concept and technology of personalized recommendation and establish own data recommendation engine in the filed of archives. The collaborative filtering technology is very suitable for building user characteristics, data characteristics and realistic construction conditions of the engine. However, it can not be applied mechanically, and should be adapted according to the actual characteristics of the application field of the file. The article designs the respective collaborative filtering matrix and the recommended implementation process respectively for the interests of the use of the set, the common use of the set of issues, the specific characteristics of the problem of file name retrieval.