论文部分内容阅读
本文介绍一个数据驱动的发现系统SYMBOL,它采用探试法的启发式策略,通过对启发式规则的假设、测试和修正,找出蕴含在两组相关符号串之间的对应规律,SYMBOL系统采用层次的知识结构,学习机制是知识层次之间的桥梁,它在系统固有的策略性知识、描述性知识,以及有关问题域的知识引导下,并在大量的实验数据的驱动下,构造一个层次结构的,关于两组相关符号串的启发式规则集。 SYMBOL系统的处理对象是两组相关符号串。这里,我们选用英语的单词和其对应的音标做为实例,也就是,SYMBOL系统在系统原有的知识引导下,在大量的实验数据的驱动下,可以发现英语单词的发音规律。
This paper introduces a data-driven discovery system SYMBOL, which uses the heuristic strategy of heuristic, through the heuristic rule assumptions, tests and corrections to find out the corresponding law contained in the string between the two groups of symbols SYMBOL system The level of knowledge structure, learning mechanism is a bridge between knowledge levels, it is in the system of strategic knowledge, descriptive knowledge, and knowledge of the problem domain, and driven by a large number of experimental data to construct a hierarchy Structure, heuristic rule set for two groups of related symbol strings. The SYMBOL system handles two groups of related symbol strings. Here, we use English words and their corresponding phonetic symbols as examples, that is, under the guidance of the original system knowledge, the SYMBOL system can find the pronunciation rules of English words driven by a large amount of experimental data.