Heart-Disease-Prediction

App Link

Our Web API looks something like this

Table of Contents

1. Dataset

2. Data Exploration

3. Data Preprocessing

4. Train-Test Splitting

5. Predictive Modelling

6. Conclusion

7. App Deployment

Dataset

• Dataset Imported from kaggle
• The Dataset contains 14 attributes

1. Age : Age of the patient
2. Sex : Sex of the patient
3. cp : Chest Pain type chest pain type
   • Value 1: typical angina
   • Value 2: atypical angina
   • Value 3: non-anginal pain
   • Value 4: asymptomatic
4. trestbps : resting blood pressure (in mm Hg)
5. chol : cholestoral in mg/dl fetched via BMI sensor
6. fbs : (fasting blood sugar > 120 mg/dl) (1 = true; 0 = false)
7. rest_ecg : resting electrocardiographic results
   • Value 0: normal
   • Value 1: having ST-T wave abnormality (T wave inversions and/or ST elevation or depression of > 0.05 mV)
   • Value 2: showing probable or definite left ventricular hypertrophy by Estes' criteria
8. thalach : maximum heart rate achieved
9. exang: exercise induced angina (1 = yes; 0 = no)
10. oldpeak: ST depression induced by exercise relative to rest OR Patient's old peak history recorded
11. slope: the slope of the peak exercise ST segment
    • Value 1: upsloping
    • Value 2: flat
    • Value 3: downsloping
12. ca: the number of major blood vessels with a fluorescent color (0-4). Fluorescent color is mainly associated with diabetes
13. thal: Thalassemia
14. target : 0= less chance of heart attack 1= more chance of heart attack

Data Exploration

• Dataset had 14 features and 303 samples
  
  
• Dataset had 5 numerical and 8 categorical features
• Slightly imbalanced dataset with 54.5% diseased patients
  
  
• No null value in dataset
• Many outliers present in our dataset

Data Preprocessing

• 1 duplicate row was present which was removed
• No null value in dataset
• After removal of outliers dataset contained 283 samples
• Applied dummy encoding on dataset

Train-Test Splitting

• Applied Standard scaler method for scaling (Standardization)
• Also applied MinMaxScaler but did not used with Application
• Test size was 20% of dataset

Predictive Modelling

Applied various Machine learning algorithms like-

• Logistic Regression
• Decision Tree Classifier
• Random Forest Classifier
• Naive Bayes' Classifier
• Support Vector Machine
• K-Nearest Neighbours
• Gradient Boosting
• Extreme Gradient Boosting (XG Boost)

1. Created Confusion matrix for all classifiers
2. Drawn plot of Roc Curves
3. Made Classification report
4. Applied RandomizedSearchCV and RepeatedStratifiedKfold for hyperparameters tuning

Conclusion

Clearly Logistic Regression algorithm worked best for our Dataset providing accuracy of 89.5%

App Creation and Deployment

• App was created using FLASK - a micro web framework in python
• App was deployed using Heroku - providing platform as a service (PaaS) that enables developers to build, run, and operate applications entirely in the cloud.
• App link - https://heart--disease--predictions.herokuapp.com/

References for developing flask api and html

• https://www.geeksforgeeks.org/deploy-machine-learning-model-using-flask/ (Full Template)
• https://www.digitalocean.com/community/tutorials/how-to-use-and-validate-web-forms-with-flask-wtf
• https://www.digitalocean.com/community/tutorials/how-to-serve-flask-applications-with-gunicorn-and-nginx-on-ubuntu-20-04
• https://www.digitalocean.com/community/tutorials/how-to-create-your-first-web-application-using-flask-and-python-3
• https://www.digitalocean.com/community/tutorials/how-to-use-templates-in-a-flask-application
• https://www.digitalocean.com/community/tutorial_series/how-to-build-a-website-with-html
• https://www.digitalocean.com/community/tutorials/how-to-use-web-forms-in-a-flask-application