Performance Analysis of Different Classification Algorithms For Heart Disease Prediction

• Get a hold of the packages and libraries that are needed for the project from here
• This project compares 7 supervised machine learning algorithms (classification algorithms) for the heart disease dataset and gives the best performed classification algorithm for heart disease prediction.
• The type of heart disease we see here is known as Cardiovascular disease (CVD)
• The flow of the project is given below

coming soon

Abstract

• Machine learning technologies have been proven to provide the best solutions to healthcare problems and biomedical communities.It also helps in the early prediction of the disease.
• Symptoms of the disease can be controlled, and proper treatment of the disease can be done due to the early prediction of the disease.
• The number of deaths due to heart attacks is increasing exponentially. Thus, machine learning approaches can be used in the early prediction of heart disease.
• Different supervised machine-learning techniques like K-Nearest Neighbors Naive Bayes Support Vector Machines Neural Networks Random Forest Classifier Decision Tree Classifier and Gradient Boosting Classifier are used for predicting heart disease using a dataset that was collected from Kaggle
• Among all other supervised classifiers, the results depict that the Gradient Boosting Classifier was better in terms of performance metrics like accuracy, precision, and sensitivity.

Important

Keywords: Heart disease, Machine learning, University of California, Irvine (UCI) Machine Repository, K-Nearest neighbors, Naive Bayes, support vector machines, neural networks, Random Forest Classifier, Decision Tree Classifier, Gradient Boosting Classifier.

How to run the Streamlit app

• Make sure to have a python interpreter with version ~=3.11
• After cloning the repo by git/github desktop open the respective IDE

pip install -r requirements.txt

• After entering the above command hit enter, which then installs streamlit seaborn pandas matplotlib scikit-learn with appropriate versions as mentioned in the requirements.txt
• In order to run the web app,

streamlit run app/navigation.py

• It might automatically open the web app directly on to your default browser or open the localhost by typing, localhost:8501

• In order to get the analysis part, upload the dataset from downloading it from the below link: Cardio Vascular Disease Dataset

• Gain knowledge on basic ML:

  • Supervised Learning

• There are two versions of this site:

• Firstly, the static type which works only on one dataset (linked above): performance-analyzer-static

• Secondly, the generic type which can intake any heart disease dataset: performance-analyzer-generic

• Created and maintained by AVidhanR