MyoCast 🫀
Predicting cardiovascular risks before the next beat. An end-to-end machine learning application forecasting myocardial infarction with 87% accuracy.
MyoCast is an intelligent predictive framework designed to assist in the early detection of myocardial infarction (heart attacks). By leveraging a pipeline of classic machine learning classifiers, MyoCast analyzes key patient physiological metrics to evaluate cardiac risk with high precision and reliability.
Rather than relying on a single algorithm, this project evaluates a suite of predictive models to find the most robust decision boundary for cardiac health classification.
- 87% Classification Accuracy: Rigorously trained and evaluated using optimized hyperparameters.
- Multi-Model Pipeline: Built with standard preprocessing (
StandardScaler) and evaluates multiple classification algorithms (KNN, SVM, Decision Trees, Naive Bayes, and Logistic Regression). - Developer-Friendly Interface: Structured, clean code written in Python using
scikit-learnfor rapid deployment and testing.
To ensure the model is both highly accurate and clinically relevant, performance was evaluated using multiple metrics (Precision, Recall, and F1-Score).
An 87% accuracy represents a highly viable, generalized model optimized to balance false positives and false negatives.
| Machine Learning Model | Accuracy | F1-Score | Status |
|---|---|---|---|
| MyoCast (Best Classifier) | 87% | 0.86 | Active Production |
| Support Vector Classifier (SVC) | 84% | 0.83 | Evaluated |
| K-Nearest Neighbors (KNN) | 81% | 0.80 | Evaluated |
| Decision Tree Classifier | 80% | 0.79 | Evaluated |
| Naive Bayes (GaussianNB) | 79% | 0.78 | Evaluated |
MyoCast is built purely in Python, utilizing standard, industry-grade scientific libraries:
- Language: Python 3.x
- Core ML Library:
scikit-learn - Data Manipulation:
pandas,numpy
To run MyoCast locally, follow these steps:
git clone [https://github.com/YOUR_GITHUB_USERNAME/MyoCast.git](https://github.com/YOUR_GITHUB_USERNAME/MyoCast.git)
cd MyoCast