The Research And Design Of Diagnostic System Based On Artificial Intelligence

Diagnostics is the application of medical knowledge to clinical symptoms to makereasonable diagnosis of illness. Clinical symptoms, including symptoms, signs, experimentand various kinds of examination, are the important materials for the attending doctors todiagnose the illness. Our country has a small number of great healers, who basicallyconcentrate in a few big hospitals in the center cities. They need to face the patients from allover the country, and the common patients need to queue up for weeks or even months toreceive treatment. For those hospitals in the remote areas, the great healers that can really becalled "experts" are rare, or even none, which leads to the high misdiagnosis rate of difficultand complicated diseases in these hospitals. In some small hospitals with poor medicalconditions, there is often the misdiagnosis of the serious illness as the small one, whichdelays the best treatment time of the patients, and even endangers the patients’ lives. Theintelligent medical diagnostic system studied in this paper is developed for this requirement.Intelligent medical diagnostic system comprises of the diagnostic system of traditionalChinese medicine, western medicine diagnosis system and the medical record storage. Thediagnostic system of traditional Chinese medicine and the western medicine diagnosis systemare the main parts of the system. In the system, the traditional Chinese medicine diagnosticsystem is based on case-based reasoning model, which uses the human informationacquisition equipment to simulate the process of TCM diagnosis and realize the TCMdiagnosis engineering, while western medicine diagnostic system, based on the neuralnetwork model optimized by the genetic algorithm, is the treatment method that obtains thedisease information from the medical record repository in the hospital information system.There are mainly four aspects of work in this subject..1. According to the TCM’s characteristic of the lack of inspection machine, this studyadopts the method of human information acquisition, to simulate the TCM’s pulsing, etc.,which achieves the use of traditional Chinese medicine way for inspection, and quantitativelysaves the inspection results to the medical record database. At the same time, on the basis ofthe characteristics of Chinese medicine, case reasoning has been set as the model to establish a TCM diagnosis expert system, which uses the growing medical records as the knowledgedatabase for reasoning to achieve the TCM diagnosis function.2. Western medicine has a lot of diagnostic methods, and the case library in eachwestern hospital has been close to being perfect. Besides, the examinations of westernmedicine are presented quantitative. Therefore, this article has designed a diagnosis modelbased on neural network, which makes full use of the existing hospital diagnosis system, andthe existing hospital case library as the network training sample, in order to achieve thewestern medicine diagnosis function.3. It takes a long time to train samples of neural network, and it is also very easy to fallinto the problem of a minimum part. This paper adopts the genetic algorithm to optimize theBP algorithm, to speed up its convergence speed and to avoid it falling into the minimum part,and specifies the appropriate scope of network structure and initial weights, so as to achievethe purpose of optimizing the whole disease diagnosis model. In this way, the diseasesdiagnosis model can be expected to turn out satisfactory results.4. The user interface is the foundation of user interaction with the system. A good userinterface can greatly decrease the difficulty of the use of the system, and improve the systemavailability and utilization. This study has designed a user interface that is suitable forintelligent diagnosis system, which establishes the means of communication between the userand the diagnosis and treatment system by the wizard mode, and thus greatly improves theease of use of the diagnosis system.