【摘要】 阈值调整是自适应文本过滤中非常重要的步骤.自适应文本过滤中训练正例少以及过滤过程中数据稀疏等问题,使得自适应文本过滤的阈值调整尤为困难.本文提出了一种阈值调整算法.该算法充分利用过滤过程中的精确率、平均相似度等观测信息调整阈值;具有使用正例少、调整速度快、系统性能高等优点.更多还原

【Abstract】 Threshold adjusting is every important in adaptive text filtering. It is also very difficult for the reason that there are few positive samples for training and sparse data while filtering. In this paper, a novel algorithm for threshold adjusting is presented. The algorithm makes use of the information such as precision and average similarity to adjust the threshold. This algorithm has the merits of using few positive samples, adjusting fast and high performance.更多还原