科技政策术语自动识别技术初探

在对科技政策领域术语的特点分析基础上，提出一种适用于科技政策领域的术语识别方法，即结合科技政 策术语的语言特点，采用统计计算的方法进行两次术语过滤过程，实现科技政策术语的自动识别。实验结果表明，本 文提出的基于科技政策术语语言特点和统计计算相结合的科技政策术语自动识别的方法具有一定的可行性，将用于科 技政策词典的构建和科技政策文本内容的深层次语义分析。

Research on Automatic Recognition Technology of Science and Technology Policy Term

The paper proposed an automatic recognition method based on characteristics and statistical computing of term. The method fully combined language characteristics and statistical information of terms. And the method of statistical calculation is adopted to carry out the two filtering process of terms. Experimental results showed that the proposed method had certain feasibility. It will have certain application value. In the next step, the method will be used for constructing the dictionary and deep semantic analysis of science and technology policy.