Research On Hypertension Diagnosis And Treatment System Based On Association Rules And Ontology

While the prevalence of hypertension increases year by year, the rate of awareness and the rate of treatment for hypertensive patients still remain at a low level, which make it quite difficult to control the hypertension effectively. Presently, the diagnosis and treatment of hypertension is still largely done manually by general practitioner, the diagnosis and treatment of hypertension which is based on artificial intelligence and computer technology is still under the way. With the extensive use of electronic prescriptions, medical institutions have accumulated a large number of electronic prescriptions of hypertensive patients. It plays a great significant for assisting the diagnosis and treatment of hypertension research to find out how to make better use of these electronic prescriptions with the help of artificial intelligence.Association Rule can be used to find out dependencies or interrelated relations among data items. And the ontology can not only display the constraints or relations between concepts, but also has a strong reasoning ability. The thesis is based on the electronic prescriptions of hypertensive patients, using association rules technology to find hidden meaningful relations from electronic prescriptions, introducing the ontology concept to describe the concepts, relations and patient examples involved in the electronic prescriptions, thereby establishing the hypertension ontology database and the SWRL rule base library. The main contribution of this thesis includes the work as follows:Firstly, in the thesis, the author analyzes the advantages and problems of electronic prescription and makes a conclusion that electronic prescriptions can certainly be widely used with the further development of the medical treatment and hygiene information construction.Secondly, the author makes a comprehensive study of the theory of association rules, analyzing the feasibility of its application in electronic prescriptions, obtaining that there exists sequence features similar to the shopping cart problem in electronic prescriptions of hypertension. With the help of Apriori algorithm, meaningful relations can be found out from the records of electronic prescription of hypertension.Thirdly, the author collects more than ten thousand electronic prescription records and researches into the methods, which is put forward by Thusitha and the other experts, to identify the patients who don't adhere on medication. In this way, the patients who don't adhere on medication can be found out and also the relationship between non-adherence to medication and blood pressure will be discovered by the Apriori algorithm. It is pointed out that combined scheme of anti-hypertension drugs and adverse drug reactions can be found by using the Association Rule.Fourthly, the construction and reasoning methods of domain ontology and SWRL rules is researched and important concepts in the electronic prescriptions about hypertension are made clear and definite in this thesis. The author combines the relationship found from the electronic prescription with other important knowledge in this field, and builds ontology library and the SWRL rules in the field of hypertension by using Protege 3.4 beta tools for the development of ontology, where he reuses the disease classification coding and drug classification instance in the ontology developed by Thusitha and the other experts. In the meanwhile, he puts forward the systemic framework of diagnosis and treatment about hypertension, and describes the process of the diagnosis and treatment in detail.The hypertension diagnosis and treatment system framework put forward in this thesis aims to aid diagnosis and treatment of hypertension. This thesis can be regarded as the exploration and attempt of diagnosis and treatment of hypertension using artificial intelligence and computer technology, hoping that it will benefit the follow-up researchers.