Fuzzy Logic Based Formalisms for Urology Diseases Diagnosis Using Cosine Amplitude Method & Gower Coefficient

Abstract

This research investigates the diagnostic use of the Gower Coefficient and the Cosine Amplitude Method (CAM) for urological disorders, with a particular emphasis on prostate cancer. The effectiveness of these techniques was evaluated using a large dataset that was obtained from Kaggle and included clinical and diagnostic information for more than 100 patients. The dataset contains a number of attributes, including diagnostic outcomes, perimeter, texture, and radius. Preparing the dataset for analysis via cleaning, normalization, encoding, and feature selection was known as data preparation. Patient feature vector similarity was calculated using the CAM, and patient similarity was handled and computed using the Gower Coefficient for a variety of data formats. The results of the model performance assessment showed that the Cosine Similarity model performed better than the others, showing the best classification capabilities and the fewest misclassifications with an accuracy of 0.98, precision of 0.9531, recall of 0.98, and an F1-Score of 0.9683. The Naive Bayes model trailed after with somewhat worse metrics, but it was still rather effective. As an example, the models for Support Vector Machine (SVM) and Kappa Coefficient performed somewhat worse, with the Kappa Coefficient model obtaining the lowest metrics. The results highlight how effective the Cosine Amplitude Method is in making precise diagnosis with little misclassification of urological illnesses. The research sheds light on the use of CAM and the Gower Coefficient in the diagnosis of urological diseases and emphasizes the need of further investigation into these techniques in clinical settings.