Research And Implementation Of Lung Cancer Diagnosis Web-based Expert System With Theory Of Certainty Factors

With the development of computer artificial intelligence and the deep research of knowledge engineering,expert system have derived a lot of related fields,such as knowledge representation,knowledge acquisition,inference method,non-determinacy and fuzzy processing,expert systems development tools and platform research.Although there are many expert systems,but for the expert systems based on the web,there has been rarely available on the Internet.At this stage,the incidence and mortality of lung cancer are growing,and most of the clinical diagnosis of lung cancer has been advanced lung cancer.The prevention and treatment of lung cancer need early detection and treatment.For the purpose of early diagnosis,ordinary people should check-up every 4-6 months in the hospital.However,such prevention and control measures are very difficult to achieve.Therefore,we can have the aid of expert systems to simulate the doctor's early diagnosis of lung cancer.The expert systems designed for the B / S mode,allow users to directly access the system through the browser for early diagnosis of lung cancer,acquire their own disease probability,provide the basis for the next diagnosis and treatment.In the theoretical study,according to the characteristics of lung cancer diagnosis,this paper provides an improved model based on deterministic factor theory of uncertainty processing method in MYCIN,which is a classical expert system,aiming at the characteristics of lung cancer diagnosis.The system model includes the transformation of the symptom qualitative probability and the quantitative probability given to the suspected patient,the quantitative probability and the conversion of the confidence in the Deterministic Factor Theory,the segmentation of the rule confidence,the weighted "AND" field processing,regular resolution "OR" field processing,and synthesizing ultimately the top hypothesis of confidence and lung cancer patients with suspected smoking information,calculating the probability of illness.The core part of the system based on the JESS expert system shell,with uses the Rete pattern matching algorithm,compared with the Markov algorithm,the system is highly efficient reasoning.In the technical research,JESS expert system shell based on the Java language,has good portability and scalability.In general,system uses the MVC design pattern,and applies Java Web development technology.Front end uses JSP web interaction,and the server uses the servlet to call the underlying Java Bean and DAO to complete the database and knowledge base access,starting the reasoning engine to obtain the results of reasoning diagnosis,and return the results to the client in the form of a diagnostic report.For the database connection,the system defines the database connection pool through the Tomcat data source,and use JNDI to find the way to connect,to improve the system operating efficiency.Finally,this paper gives the operation of the core part of the system,the system is running well.