| In the medical field, the biochemical analyzer as a kind of the conventional medicaldiagnostic facility, it includes the electronic, mechanical, software and biologicalchemistry and many other technologies. The error analysis mainly relies on the manualwhich cost a lot of time. For the variety multiple seniors of the analyzer, the multi-sourceinformation fusion technology can be used to analysis the fault. As the two kinds ofartificial intelligent method, the neural network and expert system is used frequently andwidely in the medical field.Even though the neural network and expert system is a two very effective intelligentfusion method, but also has its own disadvantages. The neural network has the conclusionreason back there box problems, as well as the expert system knowledge narrow steps, thecombinatorial explosion and knowledge acquisition bottleneck problem.In view of the above problems, a combination of the two, can effectively fosterstrengths and circumvent weaknesses, making the system more stable performance andsuperior. The paper takes the artificial intelligence method of the multi-source informationfusion technology to construct the neural network expert system model, which has beenused on the fault analysis of the automatic biochemical analyzer. It is the biochemicalanalyzer fault analysis system which based on the information fusion. At first, according tothe input knowledge of the domain expert input knowledge and the fault samples, thesystem creates the knowledge base and the system net and gets the summary, analysis andextraction to the multi-sensor data of the analyzer in a fault running process. Then the netof the system will solve the data, and send the calculate result to the expert system. Finally,the expert system will speculate the appearances of the fault, and get the correspondingreason of the fault. Both the system and the method are feasible for the accuracy of thesystem reasoning in the paper. |
PDF Full Text Request