Now days, Heart disease is the most common disease. But, unfortunately the treatment of heart disease is somewhat costly that is not affordable by common man. Hence, we can reduce this problem in some amount just by predicting heart disease before it becomes dangerous using Heart Disease Prediction System Using Machine Learning and Data mining. If we can find out heart disease problem in early stages then It becomes very helpful for treatment. Machine Learning and Data Mining techniques are used for the construction of Heart Disease Prediction System. In healthcare biomedical field, there is large use of heath care data in the form of text, images, etc but, that data is hardly visited and is not mined. So, we can avoid this problem by introducing Heart Disease Prediction System. This system will help us reduce the costs and to enhance the quality treatment of heart patients. This system can able to identify complex problems and can able to take intelligent medical decisions. The system can predict likelihood of patients of getting heart problems by their profiles such as blood pressure, age, sex, cholesterol and blood sugar. Also, the performance will be compared by calculation of confusion matrix. This can help to calculate accuracy, precision, and recall. The overall system provides high performance and better accuracy.
Heart Disease Prediction System Using Machine Learning and Data mining consists of training dataset and user input as the test dataset. Weka data mining tool with api is used to implement the heart disease prediction system. The source code of Weka is in java. The system is designed with java swing and use Weka api to call the different methods of Weka. The components used are instances, different classifiers and methods for evaluation. Supervised learning method is used here. A supervised learning algorithm analyses the training data and deduces a function from the labeled training set. It can be used for mapping new examples. The training data obtained from Cleveland heart disease database is the training example. This training data consist of the class label and its corresponding value. Naive Bayesian, J48 and Random Forest classifiers are supervised learning algorithms. They learn from the provided training examples. When a new instance with same attributes as in training data with different values other than those in the training example comes, these algorithms correctly classify the new instance based on the generalization created from the training set. Naive Bayesian, J48 and Random Forest classifiers are classify the new observation into two categories on the basis of training dataset. The training dataset is in the ARFF format. The training set consists of 14 attributes including the class attribute. Heart disease prediction system accepts input from the user through a graphical user interface.