An application of hybrid weighted similarity measure of neutrosophic set in medical diagnosis

Abstract

The study introduces a hybrid weighted similarity measure (HWSM) for the analysis of symptoms and diseases in patients using a neutrosophic set (NS). NS proves valuable for modeling uncertainty by accommodating contradictory and ambiguous information. The development of a similarity measure for NS information is crucial in various applications, particularly in medical diagnostics, to quantify similarity between sets. While existing literature provides various similarity measures for NS, only a limited number incorporates hybrid techniques. This study proposes a hybrid similarity measure that combines existing measures and integrates them with an entropy weight measure. To elaborate, distance- based similarity measures for NS are initially considered. Subsequently, an entropy weight measure is employed to calculate the attributes' weight of the attributes. The work includes formulating the properties of the proposed HWSM and its practical application in medical diagnosis, focusing on assessing the possibility of medical diagnoses in a patient. The study examines five symptoms which are fever, headache, stomach pain, cough, and chest pain. The HWSM is applied to analyze these symptoms across five different diseases, resulting in consistent and reliable outcomes. This research contributes to the ongoing enhancement of diagnostic tools for medical practitioners, addressing challenges associated with uncertainty in patient information.