Tumors are a very sickly disease and recorded as the number 2 killer disease in the world. This is because until now tumor still not found a drug that can really cure it. In Indonesia itself has many people who have suffered from tumors. Ignorance makes people reluctant to observe the early symptoms of tumor. In addition, the hospital also has problems with what type of tumor is most common in the community and their target of socialization to the community and hospital environment.
In this study, the topic of discussion focused on making patterns of tumor disease patients using Naive Bayes algorithm on Rapidminer tools using supporting variables of sex, age and place of tumor in the body. The output of this study is a posterior probability value of each variable with 66.76 percent accuracy. In addition there is also a density value in the form of a chart on each supporting variable.