基于受限自然语言和模块组合的代码自动生成

针对界面层自然语言理解的应用，该文设计一个基于受限自然语言和模块组合的代码自动生成系统。该系统允许用户输入，格式比较随意的中文需求语句，，把用户的输入转换为标准的C程序，再返回给用户。提出一种IMM分词算法，改进原有MM法在局部的最大匹配，减少错误切分句子的数量，提高分词的精度。

Code Automatic Generation Based on Limited Nature Language and Modular Synthesis

This paper aimings at a hot researched question-——the application of natural language comprehension in inter-face, designs a code automatic generation system based on limited nature language and modular synthesis. It allows users to input requires in Chinese which is free to some extent, . It converts the inputs to standard C program, then returns results to users. It raises d a syncopation-IMM, which improves the max-match in part of former syncopation——MM, reduces quantity of sentences by wrongly syncopated, and improves precision of syncopating.