论文部分内容阅读
开发了一个计算机软件系统，它能通过推断已往大坝事件／事故记录中所隐含的不明确的信息来得到复杂问题的解答。该系统曾被用于处理记述１９８１年国际大坝委员会公布的约１１００例大坝结构损毁事件的大型数据汇编的精选版本，并从按事故／事件类型分类的相联系规律中提取明确的结论。根据开发的软件，提出所生成的规律按３个不同层次分组：确证当前工作的规律，值得进一步研究的规律以及边际（无关或错误的）规律。相信这些结果会为进一步研究机器认识法以及深层次探讨某些由此获得的结果（即所生成的规律）开辟道路。
A computer software system has been developed that can get answers to complex questions by inferring the implicit information implicit in previous dam event / incident records. The system was used to process selected versions of a large compilation of data describing about 1,100 dam damage events reported by the International Commission on Dams in 1981 and to draw definitive conclusions from the association rules grouped by accident / event type . According to the software developed, the law of generation is proposed to be grouped into three different levels: the law confirming the current work, the law deserving further study, and the law of the marginal (irrelevant or false). Believe that these results will open the way to further study of the machine-learning method and to deep exploration of some of the results (ie, the laws that result).