Using standard and convectional approaches, multiform data sets are difficult to manage. Due to the dimensions internally linked and requiring cross-examination to obtain useful and accurate output, big data and analytic techniques are required. In this article, we have analyzed the Heart failure clinical dataset. Initially the dataset contains the unbalanced class distribution. So we have balanced the data by using SMOTE. To classify the patients died due to heart failure from patients who recovered, we have used SVM and decision tree as classification model. The SVM has achieved 0.83 and Decision tree achieved 0.79 accuracy. The limitation of the research is that, we have used very limited dataset, in future, we aim to enhance our dataset to get more comprehensive evaluation.
