Decision Support Method For Diagnosis Expert Based On Machine Diagnosis Reports

Nowadays, with the development of mechanical automation, the problem thatshould gain tremendous attention from daily maintenance and troubleshooting oflarge-scale machinery and equipment has become particularly important. Especially inthe electricity, water and other sectors are closely related with people’s daily lives.Domain experts who have specific diagnosis technique are needed for making sure thatthe large-scale machinery remains in good status.The objective of this paper is to propose a new generic model for supporting expertdecision, based on machine diagnosis reports which were recorded domain knowledgeand experience of experts using Naive Bayes, a machine learning method, to calculateand find the best marched diagnosis advice from machine current status. The result ofthis mode will help experts improve the efficiency in machine diagnosis than before.The whole method originates from Case-Based Reasoning. In order to achieve theobjective, first of all, the case base would be built; secondly initialization of report textsis able to finished by NLP method; finally Naive Bayes method is used to find the bestadvice text for experts.The result of simulation shows the effectiveness of the model for giving the bestadvice of machine diagnosis based on this similarity analysis from diagnosis reports.