Report · Presentation Slides · Source Code
Complications associated with acute myocardial infarction (AMI) are time-sensitive and potentially life-threatening. Pre-diagnosis of these complications at the time of a patient’s hospital admission can enable doctors to perform timely preventive measures, which can be beneficial during a patient’s emergency and post-recovery phases.
This project aims to provide insights on indicative risk factors for AMI-related complications that hospitals can utilise to enhance their current treatment protocols.
Our solution, NoMyocardial, is designed to assist doctors in the early detection of possible MI-related complications and pave the way for doctors to reduce the further development of these complications, thereby improving patient quality of life even after AMI.
Classification and Regression Tree (CART): obtain important variables (general risk indicators) for AMI-related complicationsLogistic Regression: further analyse each variable by their key risk indicators
Myocardial infarction complications from UCI Machine Learning Repository
- Data Set Characteristics: Multivariate
- Number of Instances:1700
- Attribute Characteristics: Real
- Number of Attributes: 124
- Associated Tasks: Classification
- Missing Values? Yes
| Atrial Fibrillation | Chronic Heart Failure | Relapse of MI | |
|---|---|---|---|
| Accuracy | 94.0% | 90.7% | 94.7% |
| False Positive Rate | 1.0% | 5.5% | 1.5% |
| False Negative Rate | 5.0% | 3.8% | 3.8% |
| Jing Hua | Eugene | Zhi Qi | Hao Fah | Vinayak |
|---|---|---|---|---|
 |
 |
 |
 |
 |