MATHEMATICAL ANALYSIS FOR THE PREDICTION OF TUMORS

Tumors are a threatening problem all over. The tumor's growth is suspicious in parts of the body and deals with incomplete information and uncertainty; a Rough set is a practical approach. Various studies of tumor growth have put forward; however, the feature pattern selection for the prognosis of different primary tumors has not been made still by the Rough set. We utilize medical histopathological primary tumor data to shows the approach for decision frameworks. This whole evaluation procedure confirms that the model's system execution efficiency for the prognosis of eighteen primary tumor types is 88.76%, with nearly 93% positive predictive values by the Rough set. In contrast, the accuracy by using SVM (Support Vector Machine) was 84.6%, the Bayesian Network classifier (BN) was 82.2%, the ANN (Artificial Neural Network) was 81.4%, and CR (Cox regression) was 72.6 %. In the paper, we analyze the patterns of different types of tumors applying the novel approach by the Rough set in unpredictable situations by medical data processing and rule-based decision analysis.