Fuzzy Set Theory and Kappa Co-Efficient Application for Prostate Cancer Diagnosis

This research investigates the use of Fuzzy Set Theory and the Kappa Co-efficient in diagnosing urological disorders, with a special focus on utilizing the prostate cancer dataset received from Kaggle. The dataset comprises medical records from 100 patients, each with 10 variables. The data underwent several preprocessing steps, including cleaning, normalization, and encoding of categorical variables. Utilizing a Fuzzy Inference System (FIS) was used in order to convert input variables into fuzzy sets using membership functions and generate fuzzy rules. These rules were incorporated into the Mamdani theoretical framework to produce fuzzy outputs, which were then defuzzified into crisp values. The evaluation of the FIS was conducted using many criteria, such as accuracy, precision, recall, and F1-score. The study compared the FIS with Naive Bayes, TensorFlow, and Keras models. The results indicated that the FIS achieved superior performance, with an accuracy of 99%, outperforming the other models. The Kappa Co-efficient was also used to assess inter-rater agreement. This research demonstrates the effectiveness of Fuzzy Set Theory in diagnosing prostate cancer and highlights its potential for broader applications in medical diagnostics.