This document presents a framework for early prediction of bladder cancer using DNA methylation analysis and feature selection methods. It uses two feature selection techniques: Correlation-based Feature Weighting (CFW) which assigns weights to features based on correlation with the class and average correlation between features; and Differential Mean Feature Selection (DMFS) which calculates the mean difference between normal and cancer samples to select features. The selected features are then used in a Naive Bayes classifier to classify samples as normal or cancer. The framework aims to reduce the number of features in DNA methylation datasets to improve early bladder cancer prediction.