ssjky6/stroke-predictor

Stroke Prediction App

Stroke Prediction using Machine Learning

Google Colaboratory

The data for training the model was obtained from here. According to the World Health Organization (WHO) stroke is the 2nd leading cause of death globally, responsible for approximately 11% of total deaths. Stroke is a blood clot or bleeds in the brain, which can make permanent damage that has an effecton mobility, cognition, sight or communication. Stroke is considered as medical urgent situation and can cause long-term neurological damage, complications and often death.

This web app is used to predict whether a patient is likely to get stroke based on the input parameters like gender, age, various diseases, and smoking status. Each row in the data provides relavant information about the patient.

The web app was built in Python using the following libraries and deployed in Streamlit environment.

• streamlit
• pandas
• numpy
• scikit-learn
• pickle
• Imbalanced Learn

We Experimented Outputs with Different Models:

• Random Forest Classifier

Metrics	Score
accuracy_score	0.9336505079079337
precision_score	0.9133935106123445
recall_score	0.9587708066581306
f1_score	0.9355322338830585

• Logistic Regression

Metrics	Score
accuracy_score	0.8145814581458146
precision_score	0.8058107772535387
recall_score	0.8309859154929577
f1_score	0.8182047402924861

• Adaboost Classifier

Metrics	Score
accuracy_score	0.9324932493249325
precision_score	0.9121951219512195
recall_score	0.9577464788732394
f1_score	0.9344159900062461

• SGD Classifier

Metrics	Score
accuracy_score	0.7985084222707985
precision_score	0.8023797206414899
recall_score	0.7943661971830986
f1_score	0.7983528503410113

• K Neighbours Classifier

Metrics	Score
accuracy_score	0.841969911276842
precision_score	0.7978628673196795
recall_score	0.917797695262484
f1_score	0.8536382041205192

• Voting Classifier

Metrics	Score
accuracy_score	0.8887745917448888
precision_score	0.855638745905479
recall_score	0.9364916773367478
f1_score	0.8942413497982639

• Neural Networks

Metrics	Score
accuracy_score	0.9261926192619262
precision_score	0.9045421423366529
recall_score	0.9536491677336748
f1_score	0.9284467713787086

Literature References

• Artificial Neural Network Application to the Stroke Prediction. Available from: https://ieeexplore.ieee.org/document/9203638
• Performance Analysis of Machine Learning Approaches in Stroke Prediction. Available from: https://ieeexplore.ieee.org/document/9297525