use Pearson’s Linear Correlation and the combination of Data Mining Algorithms simultaneously to improve prognosis of a kind of tumor in cancer patients

Nowadays, breast cancer is the most common cancer disease among women. Statistics shows a six percent increase in Iran which indicates it as a serious danger. However, its danger can be prevented increasingly by early diagnosis or prediction. By medical science progress, the way for developing of a system with the capability of prevention, prognosis and cure by using the new technologies is paved. Medical data mining tries to design a model and find relationships among risky factors to predict the condition of future patients with the aid of current data. We try to compare different data mining algorithms and combination of these algorithms to develop a new, efficient method with high accuracy and capability to perform on local data. Finally, proposed method which improves efficiency of Naive Bayes with Adaboost algorithm can predict the kind of benign or malign tumor with the 96/67% accuracies. Required data for this procedure is extracted from UCI site to diagnose the kind of tumor with 569 records and 32 variables.