Construction And Application Of Diagnostic Knowledge Base Of Basic Syndrome In Heart System Based On Soft Computing Methods

In the long-term clinical practice, ancient physicians put forwarded a variety of dialectical methods, such as syndrome differentiation with six meridians, syndrome differentiation with eight principles, syndrome differentiation with Wei, Qi, Ying and blood, syndrome differentiation with Sanjiao, syndrome differentiation with organs, syndrome differentiation with Qi, blood and body fluid, and so on. These dialectical methods intertwined and overlapped, failing to form a complete and unified system. In this background, Professor Wu Chengyu proposed the syndrome differentiation system of five zang-organs after many years of research and practice, which takes the five zang-organs as the core of the disease location, and determines the syndromes according to the classification of the disease features. The syndrome differentiation system of five zang-organs studies the basic syndromes with single disease location and disease feature first, and then deals with the complex and changeable syndromes simply by organically combining the basic syndromes, which is line with the thought of TCM, and can be applied to clinic. However, there are some factors hindering the promotion and application of this syndrome differentiation system, and impeding the development of information and intelligence of syndrome differentiation. For example, first, the symptoms are described fuzzily; second, there are nonlinear mappings from symptoms to syndrome elements (disease location, disease feature); third, there are complex combination logics from syndrome elements (disease location, disease feature) characteristics to the basic syndromes; fourth, the relationship between basic syndromes and syndrome elements (disease location, disease feature) couldn’t be ascertained easily by using the conventional mathematical statistics, data mining methods; and so on.Soft computing, also known as intelligent computing, is problem solving algorithms designed by simulating the biochemical processes of intelligent systems in nature (human perception, brain structure, evolution and immunity, etc.), and is solutions for a low price by fault tolerance of uncertain, imprecise and incomplete true value. It includes several computing models:fuzzy logic, artificial neural networks, genetic algorithms, evolutionary programming, chaos theory, and so on. These technologies coordinate each other. Therefore, soft computing can be used to study the syndrome differentiation system of five zang-organs to make up the shortcomings of traditional methods. Thus, in this article we took the heart diseases as the example and studied the diagnostic knowledge base of heart disease basic syndrome with soft computing. The whole process of the study was divided into the following four parts:Part â…  Study the relationship between heart system basic syndromes and syndrome elements (disease location, disease feature) characteristics by using cluster analysis, factor analysis and correlation analysis.1741medical records about heart diseases were analyzed, and the disease location characteristics and disease feature characteristics which could reflect the heart disease basic syndromes most were extracted, in order to prepare for the establishment of knowledge base later.Part â…¡ Establish the diagnostic knowledge base of heart disease basic syndromes. The diagnostic knowledge base of heart disease basic syndromes was established with BP neural network algorithm. The implicit knowledge was extracted automatically from the heart medical records by using the powerful self-learning function of the network. When the model converged, the knowledge was stored in the network node in the form of connection weights and thresholds, thus forming the diagnostic knowledge base of heart disease basic syndromes. With the limitation of convergence speed and mathematical efficiency of the algorithm, effectiveness and integrity of training samples, and so on, the effectiveness and integrity of knowledge base need to be further improved.Part â…¢ Establish the diagnostic knowledge base of heart disease basic syndromes. The diagnostic knowledge base of heart disease basic syndromes was established with the method of fuzzy identification. First, TCM experts were invited to give a score according to the importance of the disease location characteristics and disease feature characteristics. Second, the concordance rate of experts’advices was evaluated with the content validity analysis, and the validation values were used as the membership from the symptoms to the fuzzy sets of disease locations and disease features, eventually establishing the membership functions of the fuzzy sets of disease locations and disease features. Finally, the fuzzy diagnostic knowledge base of "symptomsâ†'disease locations and disease featuresâ†'heart disease basic syndrome" was established according to "the nearest principle" in the fuzzy recognition.Part IV Design the intelligent syndrome differentiation diagnostic system of heart disease basic syndrome which met the fuzzy diagnostic mathematical model above with the object-oriented software development techniques. This system was used to study the intelligent syndrome differentiation on heart medical records and the syndrome differentiation of heart diseases. In the study of the intelligent syndrome differentiation on heart medical records, the heart medical records were sorted and standardized, and the main symptoms and signs were extracted to import the intelligent syndrome differentiation system. The system would automatically differentiate the patients’ disease locations, disease features and syndromes. Compared with the original diagnosis, the system was proved to diagnose the heart basic syndrome well. In the study of syndrome differentiation of heart diseases,471medical records on Xiong Bi (angina pectoris) were collected to import to the system. The system would automatically differentiate the patients’ disease locations, disease features and syndromes. The results of syndrome differentiation were analyzed with statistical software, and the syndromes of Xiong Bi (angina pectoris) were summarized in the guidance of the syndrome differentiation system of five zang-organs.Overall, it is feasible and in line with the connotation requirements of TCM syndrome differentiation system that applying the soft computing to the field of TCM syndrome differentiation system to establish the diagnostic knowledge base model of heart disease basic syndrome. This dissertation explored the basic principles and laws of heart disease basic syndrome from multi-angle of theoretical principles, mathematical models and software applications, which has certain innovation. This idea can be further applied to the syndrome differentiation system of five zang-organs, and to serve for TCM clinical decision support.