Heart Disease Prediction Using Network Analysis

Cardiovascular diseases are a leading cause of mortality worldwide, emphasizing the critical need for early prediction and diagnosis to reduce mortality rates and enhance patients' quality of life. Diagnosing heart diseases requires a comprehensive and systematic approach, involving the analysis of complex, multidimensional, and time-consuming medical histories through traditional methods. This paper introduces a novel heart disease prediction system that combines network analysis techniques and machine learning algorithms to achieve higher diagnostic accuracy. The proposed approach will create a patient similarity network based on key physiological factors, such as age, blood pressure, cholesterol levels, and heart rate. Various centrality measures, including Degree centrality, Closeness centrality, Betweenness centrality, and Eigenvector centrality, will be applied to identify critical nodes within the network. Additionally, a machine learning-based classification model will filter network-based features to more accurately predict the risk of heart disease. The use of the \textit{NetworkX} library in Python ensures efficient calculations and scalability. The results reveal that incorporating network analysis into the diagnostic process has improved the predictive performance of heart disease classification models. The insights gained from this research may assist physicians in early diagnosis and targeted interventions, contributing to better patient outcomes and enhanced strategies in cardiovascular disease management within the healthcare system. This innovative approach, combining network analysis and machine learning, has the potential to revolutionize the way heart diseases are predicted and managed, ultimately leading to improved patient care and better population health outcomes.

Keywords:

Social Network Analysis , complex network analysis , health risks , Machine Learning , Data science