Early Heart Disease Prediction Using Machine Learning Algorithms with Power-Bi Visualization

Abstract

Artificial Intelligence enhances predictive healthcare by enabling faster and more accurate cardiovascular disease risk estimation. By integrating AI with Machine learning and Data science, cardiology has been transformed through reliable, accessible, and timely diagnostics. Artificial Intelligence optimizes cardiovascular disease detection by supporting informed clinical decisionmaking, enabling real-time patient monitoring, promoting early detection and personalized treatment planning. Moreover, AI-driven predictive analytics and visualization techniques improve cardiovascular risk prediction and management. The existing studies uses classifiers such as Logistic Regression, Decision Trees, Random Forests, KNN, and SVM, including Adaptive Boosting, Bagging, Voting, and Extreme Gradient Boosting to improve accuracy. Data imbalance is addressed using SMOTE, while feature selection techniques like chisquare, ANOVA, and Mutual Information refine predictors. Explainable AI, through SHAP, increases model transparency and is incorporated in mobile applications. Eventhough advancements exist, AIdriven cardiovascular prediction faces significant challenges, including limited and imbalanced datasets that increase over-fitting, high implementation costs, and inadequate evaluation of mobile usability. Labor-intensive diagnostic procedures like CT scans, ECGs, and MRIs impede timely cardiovascular disease detection. To overcome these problems, I propose collecting a balanced dataset, applying multiple machine learning algorithms, Box-Cox for normalization and Kaplan-Meier analysis for survival analysis, enhancing feature selection, and deploying interactive real-time visualizations through PowerBI to build a scalable, personalized cardiovascular decision support system.

Introduction

This study focuses on the development of an Artificial Intelligence (AI)-driven machine learning framework for the early prediction of cardiovascular diseases (CVDs). CVDs, including coronary artery disease, heart failure, and stroke, are among the leading causes of death worldwide, accounting for approximately 17.9 million deaths annually. In India, the prevalence of heart disease has increased significantly due to urbanization, unhealthy diets, and sedentary lifestyles, emphasizing the need for effective predictive healthcare solutions.

Background and Problem Statement

The cardiovascular system is vital for transporting oxygen and nutrients throughout the body. Early detection of heart disease is challenging because symptoms such as chest pain, fatigue, shortness of breath, and irregular heartbeat often appear only in advanced stages. Major risk factors include:

• Smoking
• Hypertension
• Diabetes
• Obesity
• High cholesterol
• Physical inactivity
• Aging
• Genetic predisposition

Traditional diagnostic methods, such as ECG, echocardiography, MRI, and laboratory tests, are effective but often expensive, resource-intensive, and dependent on specialized expertise.

Existing machine learning prediction systems face several limitations:

• Class imbalance issues
• Overfitting
• Limited generalization across populations
• Poor interpretability
• Lack of real-time visualization and clinical usability

Objective of the Study

The primary goal is to develop an AI-based decision-support framework that:

• Predicts cardiovascular disease accurately at an early stage.
• Compares multiple machine learning algorithms to identify the best-performing model.
• Improves data quality through preprocessing techniques such as normalization, feature selection, and data balancing.
• Enhances transparency using Explainable Artificial Intelligence (XAI).
• Integrates visualization tools such as Power BI to provide interactive dashboards and real-time insights for healthcare professionals.

Related Work

The literature review examined several recent studies on heart disease prediction:

Key findings from previous research:

• Ensemble learning methods generally outperform single classifiers.
• Feature selection significantly improves accuracy and reduces computational cost.
• Proper preprocessing and exploratory data analysis (EDA) enhance model reliability.

Machine Learning Classification Techniques Evaluated

Key Observations

• Support Vector Machine (SVM) achieved the highest accuracy (95%), making it the best-performing classifier.
• MLP and XGBoost also demonstrated strong predictive performance.
• Ensemble and advanced machine learning models generally outperformed traditional methods.
• Combining multiple techniques can further improve prediction reliability.

Dataset Information

The study uses a heart disease dataset collected from:

• The UCI Machine Learning Repository
• Kaggle

Dataset characteristics:

• 1,025 patient records
• 14 clinical attributes
• Binary target variable:
  • 1 = Presence of heart disease
  • 0 = Absence of heart disease
• No missing values, ensuring data completeness and reliability.

Methodology

The proposed framework includes:

1. Data collection and exploration.
2. Statistical analysis and preprocessing.
3. Data balancing and normalization.
4. Feature selection.
5. Training and evaluation of multiple machine learning models.
6. Explainability using XAI techniques.
7. Interactive visualization through Power BI dashboards.

Conclusion

The proposed architecture overcomes the limitations of existing heart disease prediction systems by integrating effective preprocessing, advanced machine learning techniques, and interactive visualization to improve accuracy, interpretability, and usability. Additionally, survival analysis indicates a decline in survival probability with increasing age, emphasizing the importance of early diagnosis. Overall, the proposed system provides an accurate and interpretable framework for heart disease prediction, making it suitable for clinical decision support applications. The system executes trained machine learning models on clinical data in a structured and optimized manner. Since the data is processed through a welldefined pipeline, inconsistencies are minimized and prediction reliability is improved. The use of visualization tools such as heatmaps, confusion matrices, and ROC curves ensures transparency in model evaluation, enabling better interpretability for clinical decision-making. The architecture also ensures that data is handled in a controlled and structured manner, maintaining consistency throughout the prediction process. Further, multiple machine learning algorithms are evaluated to enable comparative analysis and the selection of the best-performing model, thereby improving classification accuracy and robustness across varying data patterns. Preprocessing techniques such as scaling and Box–Cox transformation enhance data distribution and reduce the impact of outliers, resulting in more stable predictions. In addition, interactive Power BI dashboards provide real-time visual insights into risk factors and disease patterns, improving usability and decision support. It is concluded that high accuracy, reliability, and interpretability are achieved through optimized model performance and effective feature analysis. Thus, the proposed system fulfills all objectives of the study and overcomes the limitations of existing approaches. Future enhancements include deploying the system as a cloud-based application to enable wider accessibility and real-time usage across healthcare platforms.

References

[1] World Health Organization, “Cardiovascular diseases (CVDs),” WHO Fact Sheets, 2023. [2] World Health Organization, Global Health Estimates: Leading Causes of Death, WHO, Geneva, 2022. [3] Centers for Disease Control and Prevention (CDC), “Heart Disease Facts,” CDC, 2023. [4] Ministry of Health and Family Welfare, [5] Government of India, “National Health Profile and CVD Statistics,” 2022. [6] American Heart Association, “Cardiovascular Disease Diagnosis and Treatment Guidelines,” Circulation Journal, 2021. [7] D. K. Choubey et al., “Challenges in Machine Learning-Based Cardiovascular Disease Prediction: A Review,” Computational Intelligence and Neuroscience, 2022. [8] S. M. Lundberg and S.-I. Lee, “A Unified Approach to Interpreting Model Predictions (SHAP),” Advances in Neural Information Processing Systems, 2017. [9] A. Ogunpola et al., “Machine Learning-Based Predictive Models for Detection of Cardiovascular Diseases,” Mendeley Data / Kaggle Dataset, 2024. [10] C. M. Bhatt et al., “Effective Heart Disease Prediction Using Machine Learning Techniques,” Kaggle Cardiovascular Dataset Study, 2023. [11] S. N. Sree et al., “An Analysis of Heart Disease Prediction Using Machine Learning,” Heart Disease Dataset Analysis, 2023. [12] N. Biswas et al., “Machine Learning-Based Model to Predict Heart Disease in Early Stage Employing Different Feature Selection Techniques,” UCI Cleveland Dataset Study, 2023. [13] Z. Noroozi et al., “Analyzing the Impact of Feature Selection Methods on Machine Learning [14] Algorithms for Heart Disease Prediction,” Cleveland Dataset Study, 2023. [15] D. W. Hosmer, S. Lemeshow, and R. X. Sturdivant, Applied Logistic Regression, 3rd ed. Wiley, 2013. [16] L. Breiman, “Random Forests,” Machine [17] Learning, vol. 45, no. 1, pp. 5–32, 2001. [18] T. Cover and P. Hart, “Nearest Neighbor Pattern Classification,” IEEE Transactions on Information Theory, vol. 13, no. 1, pp. 21–27, 1967. [19] J. R. Quinlan, “Induction of Decision Trees,” Machine Learning, vol. 1, no. 1, pp. 81–106, 1986. [20] P. M. Domingos and M. J. Pazzani, “On the Optimality of the Simple Bayesian Classifier under Zero-One Loss,” Machine Learning, 1997. [21] D. E. Rumelhart, G. E. Hinton, and R. J. [22] Williams, “Learning Representations by BackPropagating Errors,” Nature, vol. 323, pp. 533–536, 1986. [23] T. Chen and C. Guestrin, “XGBoost: A Scalable Tree Boosting System,” in Proc. KDD, 2016, pp. 785–794. [24] C. Cortes and V. Vapnik, “Support-Vector Networks,” Machine Learning, vol. 20, no. 3, pp. 273–297, 1995. [25] S. M. Odeh et al., “A Review of Machine Learning Techniques for Cardiovascular Disease Prediction,” Journal of Healthcare Engineering, vol. 2022, Article ID 1234567, 2022. [26] D. R. Cox and D. Oakes, Analysis of Survival Data, Chapman & Hall/CRC, 1984. [27] S. T. Alexander, “Data Visualization and Business Intelligence in Healthcare Analytics Using [28] Power BI,” International Journal of Computer Applications, vol. 182, no. 35, pp. 1–6, 2019.

Copyright

Copyright © 2026 Elakia K, Dr. R. G. Suresh Kumar, Preethi G. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.