论文部分内容阅读
针对现有基于语境特征的术语相似度算法在语境模板生成和匹配过程中存在的不足,提出基于术语的句法依赖关系自动构造术语语境模板,进而通过语境模板匹配计算术语相似度的方法。该方法既能减少语境模板的生成和匹配困难,又将术语语境特征较好地保留在模板中。针对新方法提出具体的实现步骤,并选取基因工程领域实验数据对新方法和现有典型方法进行对比评测。实验证明,新方法在计算效果方面具有明显提升。
Aiming at the shortcomings of existing context similarity-based algorithms in the generation and matching of context templates, this paper proposes that terms-based syntactic dependencies automatically construct terminology context templates and context terminology matching to calculate terminology similarity method. This method can not only reduce the difficulty of generating and matching context templates, but also retain the context features of the terms well in the templates. Aiming at the new method, the specific implementation steps are put forward, and the experimental data in the field of genetic engineering are selected to compare the new method with the existing typical methods. Experiments show that the new method has obvious improvement in calculating the effect.