An Algorithmic Approach Using Fuzzy Soft Sets in Decision- Making for Patient-Specific Diagnosis

Fuzzy soft set (FSS) theory merges the strength of fuzzy sets (FS) and soft sets (SS) to effectively address uncertainty and vagueness in decision- making (DM) processes. It offers a systematic approach to address complex decision problems and is particularly valuable in today's data driven and uncertain world. There are numerous DM algorithms utilising FSSs. However, no prior work had employed multiple algorithms, provided algorithmic variation at each stage, and conducted a comparison analysis of the algorithms as an approach to a decision-making in a specific scenario. The current article is crafted to address these gaps. This study presents FSS theory-based algorithms for patient-disease diagnosis. In a hypothetical case study, a sequential foundational work had been set up by designating universal patient, symptom, and disease sets. The patient-symptom and symptom-disease relationships are represented using FSSs. The severity weights designated to symptoms are utilised to create a table detailing symptom severity. The sum-product composition is then applied to acquire the patient-disease diagnosis relation table, which is then defuzzified using the maximum membership principle (1st algorithm) to assign patients to the disease based on highest membership value. On the same patient-disease diagnosis output, the next algorithm introduced a modified approach in which three sets of values are constructed for each element. Triangular fuzzy numbers are used for defuzzification, and the disease suffered by each patient is indicated by highest average membership value in the row. The third algorithm incorporated disease severity and intensity-based weights. For defuzzification, the weighted average method is utilized, wherein the weighted average of membership values and allotted weights determines the disease suffered by each individual patient. The fourth algorithm identified the maximum membership value of each patient and used mean of maxima for defuzzification. We, then conducted a comparative analysis of the algorithms. The outcome revealed that there are certain similarities in the results across all four algorithms. Finally, we presented the comparison of the proposed approaches with an existing model. This research paper highlights the potential of FSS theory in patient diagnosis and provides valuable insights into diverse algorithmic approaches for accurate and personal disease identification.