Treatment Of Sudden Water Pollution Based On CBR And RBR Expert System Research

With the rapid development of the economy,various environmental problems have become increasingly prominent and environmental accidents have become more diverse,especially the sudden pollution of the water environment,which has threatened people's lives and property safety.The environmental pollution expert system can improve the success rate of decision-making in response to sudden environmental pollution.Expert system refers to an intelligent computer system that can give expert knowledge and solutions to targeted problems in a specific field and solve practical problems at the expert level in this field.At present,most expert systems are based on CBR and RBR inference modes.This paper uses a combination of CBR and RBR inference strategies to provide professional and effective solutions for sudden water environmental pollution.For the pioneering work of expert system development—the establishment of a knowledge base,the establishment of a knowledge base includes knowledge acquisition and computer language representation.Knowledge acquisition mainly adopts the methods of consulting literatures,journals,book readings,web queries,expert experiences,etc.,and streamlining the acquired knowledge events.Analyze the characteristics of all events and summarize their proportions,and obtain a series of characteristics with high proportions such as pollutants and accident geographic locations.With the description of the event features,the next step is to convert the human language of the features into computer languages.There are sub-classes of different features,which are represented by different numbers from 1 to n.In the same way,different events also have their corresponding solutions,which correspond to different numbers.A complete event is made up of features and solutions.A total of 484 events were collected,of which 363 were used for training of the expert system and the rest were used for testing.Then research on the reasoning algorithms of CBR and RBR.CBR uses the Euclidean distance and K-nearest neighbor algorithm,which combines the advantages of standard Euclidean distance and K-nearest neighbor to ensure the validity of distancemeasurement and avoid the "arbitrary and arbitrary behavior" in the nearest neighbor algorithm."Bad.And in this framework,human-computer interaction interfaces are used to characterize newly occurring events.However,the European distance of CBR reasoning has the disadvantages of inactive learning and poor generalization ability,so RBR reasoning is added.RBR reasoning uses ID3,C4.5 and CART trees.The performance evaluation of the three algorithms through accuracy,precision,recall and F1-Measure shows that the CART tree is superior to ID3 and C4.5.Therefore,RBR uses CART tree.Finally,under the UBUNTU system,the design and implementation of CBR and RBR,the system fully simulates the process of human experts in analyzing and solving problems,effectively completes intelligent reasoning,and provides emergency solutions for sudden water environmental pollution incidents,relying on two different The inferences of each other are complementary to each other to arrive at the optimal solution.Through the research in this paper,it provides the support of technology and solutions in dealing with the emergent water environmental pollution incidents.