Please use this identifier to cite or link to this item:
|Title||An Ontology-Based Approach for Diagnosing and Recommending Treatments for Myasthenia Gravis Disease|
|Title in Arabic||طريقة تعتمد على الأنطولوجيا لتشخيص واقتراح علاج لمرض وهن العضلات الوبيل|
Abstract Various diseases have emerged in recent time, which were not known to our ancestors, or they have limited deployment. The diversity of these diseases led doctors to face difficulties in diagnosing these diseases, especially when they are rare and chronic such as Myasthenia Gravis (MG) disease. Additionally, patients suffer a lot before being diagnosed correctly. The purpose of this thesis is to develop an ontology-based approach that would help doctors to diagnose the Myasthenia Gravis disease and to recommend treatments and practices that may decrease the Myasthenia Gravis impact. We reviewed several approaches and ontologies that deal with diseases such as diagnoses, patient-records, clinical decision support systems and healthcare systems. We tried to reuse that ontologies, but most of it is general ontologies for several diseases and does not focusing on specific one. Because of that we find ourselves having to develop a specific ontology for the Myasthenia Gravis disease to achieve our goals because the Myasthenia Gravis is totally different from most of the diseases. The proposed approach consists of a knowledge base (ontology and instances) and several modules such as querying, reasoning, diagnosing, and recommending treatments. A system prototype is developed with web application. It receives users' inputs such as symptoms, then returns the results in the form of query results, diagnosis results or recommended treatments and practices. The user of the system (which is a doctors), can select patient's symptoms or query about the MG disease. The system would help these doctors to decide if this patient suffers from MG disease or not, then can provide a recommended treatment for this patient through the enriched knowledge base (ontology and various instances). We made a preliminary evaluation to evaluate the diagnosing accuracy by entering information about a number of persons infected with MG disease and evaluate the results. Also, we evaluate the recommending treatments according to a human expert in Brian and Neurology by comparing his recommended treatments of a patient with a doctor's prescription who treated that patient, then with the approach recommendations to that patient. Additionally, we evaluate the efficiency of the approach by comparing the processes speed with average delay of diagnosing patients. The approach achieved a rate of accuracy in the results of diagnosing the MG disease of 86.11%, a rate of accuracy in the results of the recommending treatments of 72%. These are a better result compared to those of doctors' accuracy that treat patients' cases which is 50%. The average efficiency in the diagnosing process is 0.17 seconds and in the recommending process was 40 seconds. This time does not compare at all if we knew that the average delay in the diagnosis of patients' cases was 2.46 years. Keywords: Myasthenia Gravis, MG, diagnose, disease, treatment, recommendation systems, ontology and semantic web.
|Publisher||الجامعة الإسلامية - غزة|
|Files in this item|