From Logical Inference to Decision Trees in Medical Diagnosis

The aim of decisional systems developed for medical life is to help physicians, by providing automated tools that offer a second opinion in decision-making process. This can be connected to the diagnosis phase, treatment option, patient's evolution, identification of special medical conditions (including those emphasized by medical images analysis), or other aspects that can support physicians in their decisional activity. These automated systems are based on different mechanisms that belong to artificial intelligence (AI) domain. As it is useful to have the opportunity to select the proper model for any specific problem, it is desirable to have a choice. This study starts from an already implemented decisional system for hepatitis diagnosis (a system that uses logical inference) and extends it by the use of decision trees. The goal of this research is to create a feasible and reliable tool for diagnosis.