基于动态规划算法的专有名词切分

提出了一种对专有名词字串进行自动切分的方法，该方法对大量通用形式的专有名词进行统计分析，根据专有名词用字的出现位置和出现次数来分析，并把出现概率转换为估价值。则根据中文专有名词的特点，将中文专有名词的切分问题转换为决策树求最优解问题。采用动态规划法选取最优路径，从而确保结果的整体估价值最优。以中文姓名为例，实验表明该方法具有很快的速度和很高的正确率，用10个名字的字符串进行测试正确率为99.66%以上。

Segmentation of Named Entity Based on Dynamic Programming

This paper puts forward an approach to segment Chinese name when is put together.It carries out a statistical analysis of a great number of most commonly used named entity and sets up a corpus of name according to the position and the frequency of the occurrences of the characters used in named entity, and also transforms the frequency of the occurrences into evaluation value. Based on the Chinese characteristics used in Chinese name, it changes the segmentation of Chinese name into decision process of optimality. Dynamic programming is used to choose the optimum method. AS an example, Chinese names are used in our experiments to show that this approach works at a fairly high speed and with a high precision. The precision amounts to over 99.66% when 10 names written together are tested. Moreover, the approach can be further used in Chinese words segmentation.