Heart Disease Prediction Using Machine Learning

As heart disease remains one of the leading causes of death throughout the world, early detection and accurate prediction tools are crucial. Ordered clinical data are applied in this paper to predict the presence of cardiac disease based on a 1D Convolutional Neural Network (1D CNN) model. Relevant patient information such as age, blood pressure, cholesterol, and type of pain in the chest are part of the dataset that was downloaded from a public health database. It begins with a thorough data preprocessing involving missing value handling, feature scaling, and duplicate elimination. The class unbalance was solved with the assistance of the Synthetic Minority Over-sampling Technique (SMOTE) and the fairness of predictions was improved. The data set was then converted to a format in which CNN can function. Multiple convolution layers with ReLU activation, max pooling, flattening, and Dense Layers with dropout to prevent overfitting make up the architecture of the proposed 1D CNN model. Sigmoid activation is used in the final layer for binary classification. Binary cross-entropy loss and Adam optimizer were used in training the model for 50 epochs. Accuracy, precision, recall, F1-score, and ROC curve were utilized to quantify performance. Having an accuracy level of around 85%, the model showed an outstanding capacity for identifying trends within medical data and making accurate predictions about heart disease. Visualizations like ROC plots, confusion matrices, and loss/accuracy curves also tested the effectiveness of the model. This paper exemplifies the application of deep learning, in the form of 1D CNNs, for creating efficient, scalable, and precise diagnostic tools. The findings provide evidence toward integration into real-time clinical decision support systems and further research based on larger and more diverse datasets.