Heart Disease Prediction mini project presentation

Heart Disease Prediction mini project presentation

1. My Cardio Care Early Detection for a Healthier Future Our Team : Mrunal Khade Samiksha Kshirsagar Vanshita Munot Aditya Yevate Guided By: Ms. Janhavi Kale

2. Problem Statement Heart disease remains a leading cause of death worldwide. Early detection is critical for effective treatment and improved outcomes. Solution User-Friendly Interface: Easy for individuals and healthcare providers to use. Machine Learning: Uses Gaussian Naive Bayes classifier for heart disease prediction. Health Metrics Analysis: Analyzes key metrics like age, blood pressure, cholesterol. Early Detection: Provides early warnings for timely medical intervention. Validated Predictions : Utilizes Cleveland heart disease dataset. Accessible : Makes advanced predictive analysis accessible to a wider audience. Many predictive tools are either too complex or inaccessible to the general public and smaller healthcare providers. There is a need for a simple, user-friendly tool that can provide reliable predictions based on common health metrics. MyCardioCare Application: Lack of Accessible Predictive Tools: Rising Incidence of Heart Disease:

3. Idea of Project Predictive Analysis for Heart Health Project Ideas : Develop : MyCardioCare aims to use machine learning to predict heart disease based on key health metrics. Purpose : The application focuses on early detection and intervention to improve heart health outcomes. Objectives: Accuracy: Build a precise predictive model for heart disease. User Experience: Design an intuitive interface for ease of data input and result interpretation. Dataset Utilization: Implement the Cleveland heart disease dataset for reliable predictions. Proactive Health Management: Provide a tool that supports early intervention and preventive care.

4. Methodology Data Collection Obtain Cleveland heart disease dataset Preprocess data (clean, normalize, encode) Model Development Implement Gaussian Naive Bayes Train and optimize model Model Evaluation Validate with cross-validation Test on separate dataset User Interface Design Develop user-friendly interface Implement data input and result display Integration and Testing Integrate model with interface Perform end-to-end testing Deployment Deploy application Monitor and maintain

5. USER WORK FLOW 1. Open Application 2. Login 3. Enter Health Metrics 4. Submit Data 5. Receive Prediction 6. Reset or Exit

6. Our Output (Level 0) Output Message: "PRESENCE OF DISEASE: LESS THAN 25% (NEGLIGIBLE)" (Level 2) Output Message: "PRESENCE OF DISEASE: LESS THAN 50% (NEED TO BE CAREFUL)" (Level 3) Output Message: "DISEASE...!!! NEEDS TREATMENT.. STEPPING TOWARDS HIGH RISK" (Level 4) Output Message: "BE CAREFUL...HIGH RISK!!!" . No Heart Disease Output Message: "GREAT...YOU ARE STRONG AND FIT...!!!" (Level 1) Low Risk Moderate Risk High Risk Very High Risk output will appear here

7. Operating System: Windows 10+, macOS 10.13+, Linux Programming : Python 3.6+, numpy, scikit-learn, appJar Development: PyCharm, Visual Studio Code Data: Cleveland heart disease dataset (CSV) Dependencies: pip Version Control: Git (GitHub/GitLab) Testing: pytest Deployment: PyInstaller, cx_Freeze Software Requirements

8. Why Our Model ? Aspect Model Type Accuracy User Interface Data Handling Predictive Capability Integration Training Data Existing Model Various Models Variable complex or non interactive Extensive preprocessing Limited Risk Levels Often Complex Varied and not always validated My Cardio Care Gaussian Naive Bayes (GNB) High with Cleveland data User-friendly GUI Integrated preprocessing Detailed risk levels Seamless integration Cleveland dataset Advantages of MyCardioCare Effective for small datasets and medical data. Robust accuracy in predicting heart disease. Intuitive and easy to use. Simplified and streamlined data handling. Provides detailed risk assessments. Smooth integration for better user experience. Well-validated data ensures reliability.

9. Literature Review Detrano, R., et al. "International application of a new probability algorithm for the diagnosis of coronary artery disease." The American Journal of Cardiology 64.5 (1989): 304310. This paper discusses the Cleveland heart disease dataset, which was pivotal in understanding the features and labels used in our model. Paper 1. Paper 2. John, G.H., Langley, P. "Estimating Continuous Distributions in Bayesian Classifiers." Proceedings of the Eleventh Conference on Uncertainty in Artificial Intelligence. (1995): 338-345. This paper provided detailed information on the Gaussian Naive Bayes algorithm and its applications. Paper 3. Heller, G., et al. "Analysis of the Survival in Patients with Coronary Artery Disease and Left Ventricular Dysfunction Treated with Surgical or Percutaneous Revascularization Versus Medical Therapy Alone." Journal of the American College of Cardiology 50.3 (2007): 224-230. This paper compares various treatment methods for coronary artery disease and their outcomes, highlighting the importance of early prediction and intervention.

10. Literature Review Heller, G., et al. "Analysis of the Survival in Patients with Coronary Artery Disease and Left Ventricular Dysfunction Treated with Surgical or Percutaneous Revascularization Versus Medical Therapy Alone." Journal of the American College of Cardiology 50.3 (2007): 224-230. This paper compares various treatment methods for coronary artery disease and their outcomes, highlighting the importance of early prediction and intervention. Paper 4. Paper 5. Aha, D. W., et al. "Instance-Based Learning Algorithms." Machine Learning 6.1 (1991): 37-66. This paper discusses instance-based learning algorithms and their application in medical diagnosis, offering alternative approaches to disease prediction.

11. Samiksha Kshirsagar Vanshita Munot Aditya Yevate Mrunal Khade Data Scientist Project Manager Backend Developer Work Distribution Frontend Developer

12. Thankyou