DIABETES MODEL (SVM) PREDICTION USING MACHINE LEARNING CASE STUDY: NATIONAL INSTITUTE OF DIABETES AND DIGESTIVE AND KIDNEY DISEASES (ARIZONA, USA)

Abstract

A chronic illness that affects millions of individuals worldwide, diabetes poses serious health hazards if left untreated. Improving patient outcomes and limiting serious consequences can be achieved by accurately predicting diabetes and detecting it early. This research aims to create a diabetes prediction model by applying the Support Vector Machine (SVM) algorithm, a potent supervised machine learning method well-known for its efficiency in classification applications. A dataset with numerous health-related characteristics, including blood pressure, body mass index (BMI), age, and glucose levels, was used to train and assess the model. Making a trustworthy and accurate algorithm that could identify people as either diabetics or non-diabetics based on their data was the main goal. The model was created using Python and other libraries, including Scikit-learn for machine learning tasks, Pandas and NumPy for data processing, and Matplotlib and Seaborn for data visualization. In addition, a Django web application with HTML, CSS, Bootstrap, and JavaScript was created to offer an interactive user interface for data entry and prediction viewing. The backend database utilized to store user data and prediction history was SQLite. The outcomes showed that the SVM model had a good accuracy rate, which made it a useful tool for clinical diabetes prediction. This paper offers a useful implementation methodology for incorporating predictive models into user-friendly online apps and highlights the possibilities of machine learning in the healthcare industry. To further enhance the accuracy and generalizability of the diabetes prediction model, future research should consider integrating additional features such as genetic data and lifestyle factors, as well as exploring more complex algorithms like deep learning models. Implementing the system in a real clinical setting could also help validate its effectiveness in practice. Additionally, expanding the scope of the study to include different population demographics could ensure that the model is applicable across diverse patient groups. Regular updates and improvements based on user feedback and advances in machine learning techniques are essential to maintaining the system's relevance and effectiveness in predicting diabetes.