A Decision Model to Predict Clinical Stage of Bladder Cancer

Abstract

An extensive variety of computational techniques and tools have been developed for information investigation in medical domain. In this article, we have exploited those accessible technological headways to predict stage of bladder cancer. Present system of examination involves an invasive procedure called “cystoscopy” to find high risk patients. This unique research work helps in determining contributing factors (demographic as well as pathological) leading to the progression of bladder cancer. The proposed predictive model if used by restorative specialists and professionals will help in eliminating unnecessary cystoscopy in patients with low stage and grade diseases. As additional contribution, this article also validates the performance of our earlier designed hybrid model dealing with missing value imputation (HPM-MI) for the diagnosis of stage of bladder cancer. This model is fit for precise prediction in nearness of vast missing values. The evaluation of this model is examined by means of the classification accuracy, precision, recall and F1 measure. The dataset is collected from Max Super Specialty Hospital (a unit of Balaji Medical and diagnostic research centre), Patparganj, Delhi, retrospectively from records of cancerous patients. The results are found to be 82.39% precise and accurate when compared with actual information.