作者: Peter Wyard
DOI: 10.1007/3-540-58473-0_151
关键词: Affix grammar 、 Grammar systems theory 、 Adaptive grammar 、 Link grammar 、 Extended Affix Grammar 、 Natural language processing 、 Generative grammar 、 Artificial intelligence 、 Mildly context-sensitive grammar formalism 、 Context-free grammar 、 Computer science
摘要: This paper describes results on the inference of two classes context free grammar (CFG), using a genetic algorithm (GA). The first class is that n-symbol palindromes, where n=2 to 4; second small natural language grammars. use different normal forms grammars was compared experimentally. encodings in chromosomes GA, and implications these representations within search are discussed. It concluded by paying attention representational issues, worthwhile may be achieved GA.