The purpose of this project is create a classification model capable of identifying individuals at risk of heart failure based on certain health-related parameters. Such models can potentially aid doctors/patients in the future by identifying certain trends and take the required precautionary measures as soon as possible. High recall is a must in such healthcare related models!
- Inferential Statistics
- Machine Learning
- Feature Engineering
- Predictive Modeling
- Deep Learning
- Data Visualization
- Classification
- Python
- Pandas, TensorFlow, SkLearn
- Collab
- This Notebook is based off an open source dataset available on www.kaggle.com where I have created models to classify patients who can potentially witness heart failure on the basis of various parameters! . The best model had an accuracy of 94% and a recall of 91%
- All models are subject to betterment with more stringent hyper-parameter tuning. This can be achieved by random selection, brute force methods, etc. Various other classifiers can also be used, but the most standard classifiers have been considered in this notebook.
- Recommend standard practices for data transformation, outlier detection, and null value substitution have been incorporated in this notebook.
- This code has been UPVOTED by 15 People, Including Kaggle Grandmasters (Highly recognised people for their achievements in the data science Community). I have received a bronze medal for my code in the community.
$ git clone git@github.com:akarshsinghh/Heart-Failure-Classification.git
$ cd Heart-Failure-Classification
$ npm install
$ npm start
NOTE: The port by default will be http://localhost:3000/
Alternatively one can simply download the notebook and dataset, open in platforms like Jupyter, Collab, and Run each cell to see results!
- Feel free to contact team leads with any questions or if you are interested in contributing!
- https://www.linkedin.com/in/akarshsinghh/