Research On Emergency Cases Reasoning Based On Knowledge Element And Fuzzy Cognitive Map

Researches on case reasoning, frome previous experience, could do help to solve the emergency case problemes and provide decision support for emergency management. Currently, there are two kinds of emergency case reasoning methods. The most mainstream of case reasoning methods is Case-Based Reasoning. Lots of researchers focus on the study of case structure and retrieval of cases. All these scholars do researches from the perspective of case attributes and case features. There are some other researchers proposed some case reasoning methods from the perspective of the substance of emergency cases. However, there are some common problems of all of these methods. Firstly, the current CBR methods which are analogous inference methods focus on the case attributes or features. To solve problem, these methods try to inquire the past cases to match the current emergency cases. They cannot discribe the emergency case structure efficiently and clearlly. Secondly, researches on case reasoning from the perspective of the substance of emergency cases are still at the stage of exploration. And the reasoning process of these methods are uncomplicated. At last, these two kinds methods focus on the case reasoning consequences and ignore the emergency cases development and the implicit knowledge between the case elements, and also the human factors plat an important role.To avoid these disadvantages, this paper presents a new methodology for modeling and doing case reasoning using machine learning and data mining techniques. This paper take the ordinary emergency cases as the object, and build a case reasoning model by discovering the related information of the essential factors and the causal information between different historical cases. In this paper, the cases are decomposed into the knowledge elements. We take the knowledge elements with field features as the essential factors to extract the causal knowledge information by using the association rule mining techniques. By learning the causality information contained in the association rules, a lot of causal information chains can be found out. To model the fuzzy cognitive map, we distinguish the causal information chains to find out the nodes and learn the weight between the nodes. In the FCM model, the nodes represent the factors of emergency cases and the weighted directed arcs between nodes represent the relationship of the two connected nodes. At last, this paper take the gas cases to verify the feasibility and the effectiveness of the model. This study acquire the knowledge elements with field features and obtain the association rules between them using data mining techniques. Taking these knowledge elements as the nodes of the FCM model, it could do gas case reasoning efficiently. Empirical evaluation on real gas emergency cases shows that our pundit algorithm performs as well as non-expert humans. This paper proposes the FCM reasoning model based on the knowledge element, the first advantage is that, from the perspective of substances of emergency cases, it could discribe the emergency cases structure and the the implicit konwledge between the case elements clearly and efficiently taking the field konwledge elements as the nodes of the FCM model, which could avoid the human factors. The second, while reasoning the case consequences, this method can display the interaction process between the elements of emergency cases, which makes the case reasoning methods more effectively and correctively. It can help decision makers to grasp the condition of the emergency cases, guide the development of the emergency events, predict the results of the emergency cases, etc.