<![CDATA[Abstrct ? To be able to maintain a minimum stock of blood transfusion, donate their blood,potential pendonror returned must be known, since the blood results tranfusi can no longer be used after 42 days. During this time in predicting potential donors donate their blood again produces different accuracy on some algorithms of classification by using dataset are different. So it is not yet known where the dataset with algorithm suitable for predictions. Need to find out and distinguish between potential blood donors donate their blood again and what not, need to be built so that the blood donor unit can take the decision to keep the blood stock to keep it secure. In this study performed comparisons of Neural Network Algorithm, K-Nearest Neighbor and Naïve Bayes data is applied to the donors and blood donor transaction data with RFMTC dataset and PMI dataset. From the test results by measuring the performance of these three algorithms when applied to both of the dataset test using the Confusion Matrix and ROC Curves, it is known that neural network algorithm with dataset RFMTC has the value of the highest accuracy. Soobtained the use if neural network algorithm with dataset RFMTC which fits in this research to applied to the prediction of potential blood donors.Keyword: Blood Donors, Neural Network, Naïve Bayes, K-NN, RFMTC.]]>
Copyrights © 2016
