Parameter selection using Shannon entropy in soft set theory: an application to medical diagnosis

We integrate Shannon entropy into soft set theory to provide a simple, interpretable parameter-ranking method for systems described by soft sets. We define an importance score R (p) =H (p) /H (P) for each parameter using individual and parametric entropies, prove basic ordering properties, and show how joint entropies and mutual-information–like measures support robustness checks. The method is illustrated on an influenza symptom dataset and compared against classical ranking approaches. Results show the approach identifies influential symptoms while remaining computationally straightforward on binary symptom data. We discuss complexity, limitations, and public availability of the dataset.