An Effective Framework for Early Detection and Classification of Cardiovascular Disease (CVD) Using Machine Learning Techniques

Cardiovascular disease (CVD) is a condition that can kill and is becoming more common around the world. CVD and other heart illnesses could be found and predicted early on, which could save many lives. This needs a careful study of clinical data. Everyone in the healthcare industry benefits from predictive machine learning algorithms’ ability to enhance a doctor’s sense of things, as this can lead to more accurate patient diagnoses and more effective treatment. Reducing and learning more about heart disease symptoms with ML is a viable option. When it comes to categorizing the data, such as whether a person has cardiac problems or not, ML algorithms are more precise and expedient. A goal of this work is to create a reliable ML model for diagnosing cardiac illness so that it may be used to help save lives. The feature selection algorithm’s Recall, accuracy, precision, and F1-score parameters and a performance assessment matrix are used to evaluate the quality of our model. All the features of the dataset and a subset of them were used to test the work that was suggested. Reducing the number of features influences the evaluation matrix and accuracy of algorithms. Utilizing a UCI set of data based on people’s medical traits, this study aims to create multiple ML techniques with the goal of improving early identification of cardiovascular illness. The goal of a UCI ML heart disease dataset is to facilitate the evaluation and comparison of various ML approaches. A Cleveland heart disease (HD) dataset is analyzed using principal component analysis (PCA) to extract useful information about the condition. Using the Gradient Boosting (GB), AdaBoost (AB), and Multi-layer Perceptron (MLP) algorithms, we were able to create a model with a prediction accuracy of 91.66 percent, 90 percent, and 90 percent, respectively, for cardiovascular disease risk. The data visualization was built to display the connection between the characteristics. We also analyzed the results of our experiments on all the algorithms we developed. Based on these findings, the Gradient Boosting method seems to be the most effective one for CVD classification and prediction. Our proposed model for disease classification and prediction is applicable to any healthcare system in the world.