Petri Net Based State Identification ? Risk Prognosis Knowledge Reasoning Model

As a subset of knowledge reasoning,health-related knowledge reasoning is charactered as large volume,complicated structure,and numerous parameters(weight,time,probability,etc).As China society aging,growing attentions have been put into health problem.Based on traditional Petri Net and its developed models,this article discussed the solution algorism to health-related knowledge network reasoning,and finally proved the availability of the algorism.According to different needs,health-related knowledge could be divided into 3 kinds: state identification,risk prognosis and solution management.At first we designed a plan to state identification knowledge,in terms of its uncertainty,we used Fuzzy Petri Net as model basis.In the end,we used an extract of <China guidelines of prevention and treatment to hypertension> to validate the effectiveness of algorism.Unlike state identification,while dealing with risk prognosis knowledge,it was uneasy to solve the probability and time-delay parameters in the original rule database.To solve such problem,we used an improved probability & time delay Petri Net.By giving probability and timing function to Transitions,we achieved a reasonable model of risk prognosis.And proved the effectiveness by giving modeling example of <Framingham CHD risk score>.As a solution to the low efficiency and high mistakable problem while modelling large scale Petri Net(PN)by hand,this article illustrated an automatic PN modelling method in health knowledge reasoning field.After giving the main technic path and algorithm,we finally proved the achievability of such algorithm by programming.The main parts of algorithm included network reflecting,graphic elements locating and PN reasoning.As the experiment indicated,while facing large scale health knowledge reasoning,automatic PN modelling algorithm had higher efficiency,lower mistake and higher dynamic performance comparing to manual modelling method,which would high probably push the application of PN in health knowledge reasoning.