Overview / Usage

The project uses Intel OneAPI's Scikit-Learn algorithm and oneDAL library to create a machine learning model to predict cardiovascular illness. The objective is to identify those who are very susceptible to CVD and offer them specialized preventative measures to improve patient outcomes.

The initiative focuses on the issue of early detection and prevention of CVD, a major cause of death globally. Conventional risk prediction models rely on a small number of risk factors, including age, gender, and smoking status, and have a limited degree of accuracy. Contrarily, machine learning algorithms can take into account many risk factors and offer more accurate forecasts, improving healthcare providers' ability to make decisions.

To determine the best-performing model, the project develops and assesses a number of machine learning algorithms, including logistic regression, SVM, KNN, decision tree, random forest, and XGBoost. The project also makes use of the oneDAL library and Intel OneAPI's Scikit-Learn algorithm to speed up computations and enhance the efficiency of the model.

​

Healthcare professionals and researchers can use this work/research to identify high-risk patients and customise preventive measures. It can also assist in the development of targeted treatments and health policies by policymakers and public health professionals to lessen the burden of CVD. Overall, this initiative has many practical applications and has the potential to improve community health and patient outcomes.

Methodology / Approach

In this study, we are predicting cardiovascular illness using machine learning algorithms. This is a challenging topic that calls for the examination of numerous risk factors. We are using Intel OneAPI's Scikit-Learn algorithm and oneDAL package to tackle this issue in order to speed up computations and enhance model performance.

Preprocessing entails preparing the dataset for machine learning algorithms by cleaning and manipulating the data. For this assignment, we use the Scikit-Learn preprocessing module, which offers a number of tools including standardization, normalization, and imputation.

Then, we choose the relevant machine learning algorithms, including XGBoost, decision trees, random forests, SVM, KNN, and logistic regression. For this work, we use the Scikit-Learn toolkit, which offers a variety of machine learning tools and techniques for model selection and hyperparameter tweaking.

We make use of Intel OneAPI's oneDAL library, which offers machine learning task-optimized algorithms and data structures, to further enhance model performance. With the help of this library, we can parallelize computations and utilize cutting-edge hardware, including CPUs, GPUs, and FPGAs.

Using metrics like accuracy, F1 score, and AUC-ROC curve, we assess model performance. Additionally, we cross-validate our results to make sure they are reliable.

​

In conclusion, our methodology involves preprocessing the data, choosing relevant machine learning algorithms, optimising computations, and assessing model performance using the libraries Scikit-Learn and oneDAL. We can create a machine learning model for predicting cardiovascular illness that is accurate and effective thanks to these frameworks, standards, and methods.

​

Technologies Used

• Technologies: Machine learning, Data Science
• Libraries: Scikit-Learn, Pandas, Numpy, Matplotlib, Seaborn, OneDAL
• Tools: Jupyter Notebook, Git, GitHub
• Software: Python 3.7 or above, Anaconda, Intel OneAPI Base Toolkit
• Hardware: CPU, GPU, and FPGA
• Intel Technologies: Intel OneAPI's Scikit-Learn Algorithm and oneDAL library.

Repository

https://github.com/JoelJJoseph/CardioVasular-Disease-OneAPI