International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 12 Issue: 05 | May 2025
p-ISSN: 2395-0072
www.irjet.net
Predict Diabetes using Machine Learning Kalyani , Shriya Singh , Shalu Kumari, Prof. Mukesh Kumar Bhardwaj Dronacharya Group of Institutions, Greater Noida, Uttar Pradesh, India ----------------------------------------------------------------------***----------------------------------------------------------------------
Abstract - This study uses the Support Vector Machine
Indians Diabetes Database, which provides a fundamental dataset for this study and is necessary for precise forecasting.
(SVM) algorithm after comparing with other algorithms to provide a machine learning method for early diabetes identification in young people. AUC, F1 score, recall, accuracy, and precision are some of the metrics used to compare each classifier's performance. A Kaggle dataset including lifestyle, medical, and demographic characteristics is used by the system. Feature scaling and data preparation were followed by the application of SVM to create a classification model. With an accuracy of almost 92%, the experimental results show how successful the suggested model is. The system seeks to promote prompt medical action and assist young people in early risk identification.
This research will explore the complexities of putting into practice a Support Vector Machine (SVM) model, which is well-known for its resilience in classification tasks. Because the SVM algorithm finds the best hyperplane in the feature space to divide classes, it is especially well-suited for binary classification of patients with and without diabetes. The complete procedure, from feature selection and data preparation to model training and assessment, will be described in the paper. We will evaluate the SVM model's predictive power for diabetes using a range of performance indicators. This study ultimately seeks to support current initiatives to use machine learning to improve healthcare outcomes by offering a foundation for upcoming developments in diabetes management and prediction. In order to forecast diabetes using normal medical data, this work investigates a cost-effective machine learning-based approach that uses the Support Vector Machine (SVM).
Keywords — Diabetes Prediction, Machine Learning, SVM, Young Generation, Classification, Health Informatics.
1. INTRODUCTION High blood sugar levels are a hallmark of diabetes, a chronic metabolic disease that, if left untreated, can cause serious health problems. Early detection and intervention are essential for enhancing patient outcomes due to the rising incidence of diabetes worldwide. In the medical field, machine learning (ML) has become a potent instrument, especially for estimating the risk of diabetes. The Support Vector Machine (SVM) algorithm is unique among machine learning algorithms because it can effectively handle high- dimensional data and classify intricate patterns. With an emphasis on methodology, data preparation, feature selection, and model evaluation, this research study investigates the creation of a diabetes prediction system using SVM. In order to enable prompt medical intervention and individualized treatment programs, the proposed method seeks to precisely identify those who are at risk of acquiring diabetes by utilizing past patient data. Incorporating SVM into diabetes prediction not only improves accuracy but also advances the field of predictive analytics in healthcare, opening the door for creative ways to tackle this expanding problem.
2. LITERATURE REVIEW Recent years have seen a considerable increase in interest in the use of machine learning (ML) approaches for diabetes prediction, with numerous studies investigating various algorithms to improve prediction accuracy. This field has made extensive use of conventional techniques like Decision Trees, K-Nearest Neighbor (KNN), and Logistic Regression. In diabetes prediction, for example, decision trees are a common choice for preliminary research due to their interpretability and simplicity of usage (Quinlan, 1986). KNN is renowned for its ease of use and has been successfully applied, especially in smaller datasets, where it can produce excellent outcomes (Cover & Hart, 1967). A statistical technique called logistic regression has been used extensively in medical research for binary classification tasks because it offers a baseline against which more intricate models may be compared (Hosmer & Lemeshow, 2000). These conventional techniques, however, frequently falter when dealing with high- dimensional data, which is typical of medical datasets. On the other hand, Support Vector Machines (SVM) have become a strong substitute, especially for high-dimensional binary classification applications. SVM works by determining the best hyperplane to
The importance of diabetes prediction systems cannot be emphasized because they are essential to public health because they allow for early illness detection and treatment. Important health markers like age, body mass index (BMI), and glucose levels are included in the Pima
© 2025, IRJET
|
Impact Factor value: 8.315
|
ISO 9001:2008 Certified Journal
|
Page 837