Abstract: Instead of traditional method of inputting keywords according to human experience, this paper analyzes a new information retrieval method. When searching special area information, it fuses sample documents, extracts common keywords, and adjusts the sample documents to control the search space and result sum. The method avoids the warp being at the traditional method inducing by the subjectivity and unilateralism of the human experience to a certain extent, can reach the excellent harmony between the recall ration and the result sum, achieving the best cost-performance.

Key words: text filtering, samples fusion, similarity computation, intervals control

摘要： 探索了一种新的检索方式，代替了依赖人工经验输入关键字检索的传统方法。在检索特定领域信息时，通过相关样本集融合，提取出关键词集，通过调节样本集实现关键词集的柔性控制，以调控搜索空间与结果取向。该方法在一定程度上避免了人类经验的主观性、片面性和关键词任选偏差，可以使查全率与结果数量达到最佳协调，实现最优性价比。

关键词: 文本过滤, 样本融合, 相似度计算, 区间调节

CLC Number: 

• TP18