Developing A Knowledge Based Clinical Laboratory Recommender System Using Rule Based Approach

Abstract

This work presents a rule based clinical laboratory recommender system. The major goal of this work is to create a rule-based clinical laboratory recommender system. The growing demands for key healthcare resources such as clinical expertise and facilities have motivated the emergence of Artificial Intelligence (AI) based decision support systems. These systems will be vital in cutting down on the time it takes doctors to diagnose illnesses and pointing them in the direction of the most accurate diagnostic techniques. Increasing the number of patients seen by each doctor each day, especially in places where there aren't enough physicians. To achieve this objective, knowledge was acquired using both structured and unstructured interviews with ten experts, which was selected purposely from Jimma University Referral Hospital and Odaa Hullee Hospital. The main goal of this study is to create knowledge-based clinical laboratory recommender systems that use a rule-based methodology. In addition to that, knowledge is acquired from secondary sources (internet, articles, manuals and some reports). The acquired knowledge was modeled using decision tree to represent with concepts of sign, symptom, risk factors, physical and laboratory tests involved in diseases diagnoses purpose. The study was used rule based reasoning method through backward reasoning approach for diagnoses diseases, and the prototype was developed with swi Prolog. The conceptual model of the knowledge-based system made use of a decision tree structure, which makes it simple to comprehend and evaluate the steps involved in patient diagnoses for clinical laboratory recommender systems. The prototype is created with lpa win prologue utilizing "if-then" rules based on the conceptual model. Backward chaining is a technique used by the prototype to infer the rules and offer suitable recommendations. The system evaluators have generally given the prototype knowledge based system's performance positive feedback.89% of the users of the prototype, according to the system assessors, are happy with it. Additionally, the system's performance is assessed utilizing predictive validation methods with five test scenarios. The prototype's accuracy is roughly 90%, according to the validation test case findings. The integration of rule-based and case-based reasoning is recommended to enhance the performance of the inference engine. It is also suggested that knowledge-based systems be developed in various local languages so that users can communicate with each other in their tongues and that the research in the relevant field is strengthened.