One of the main causes of death for women in developed countries is now recognized to be breast cancer. The best method for reducing mortality rates is early breast cancer identification. The capacity to detect breast cancer early, however, is necessary for faster therapy. In this method, feature acquisition is carried out during the pre-processing stage by applying Dyadic Transformation. The training data's column consistency patterns are then discovered using the useful technology of biclustering mining. The recurring patterns in tumors with the same label could serve as a possible diagnostic guide. Then, using a novel way for combining rules, the classification problem in various feature spaces is resolved by constructing component classifiers of the AdaBoost algorithm utilizing the diagnostic rules (PCDFS).