Analysis of Blood Transfusion Dataset Using Data Mining Techniques

Data mining is the process of examining and learning from massive dataset. Here, the Blood Transfusion dataset is being processed, and classification algorithms were used to obtain the necessary knowledge. First, five classification algorithms of bagging, LogitBoost, J48, ClassificationViaRegression and RandomForest are applied to the blood transfusion dataset. Before normalization, the values for the performance metrics of accuracy, sensitivity, specificity and error rate were found for the above classification algorithms. Then, the dataset was normalized within the values between 0.0 and 1.0 and the same performance metric values were found with the above classification algorithms. From the results of the performance metrics, RandomForest was the best classification algorithm for the blood transfusion dataset.