Logo image
Back
Diabetes disease prediction using machine learning
Journal article

Diabetes disease prediction using machine learning

Jobin M. Scaria, Haritha Bargav, Sharmila Joseph, Sasikumar V R and A Baseera
IBSUniversity Journal of Business and Research, Vol.2(1), pp.18-31
06/2025
url
Diabetes disease prediction using machine learningView
Published (Version of record) Free to Read

Related links

Metrics

1 Record Views

Abstract

Diabetes prediction mellitus jinja Random Forest flask web application questionnaire healthcare AI
Diabetes mellitus is a chronic disease affecting millions worldwide, and early predictions play a crucial role in preventive healthcare. This project aims to develop an efficient and user-friendly system for diabetes disease prediction using machine learning techniques. Leveraging the Random Forest Classifier from the scikit -learn library, the model was trained on the 2019 diabetes dataset after proper preprocessing with Pandas and Label Encoder. Model performance was evaluated using standard metrics like accuracy score after data splitting using train_test_split. To visualize model performance, Matplotlib was used to create comparative bar charts. The project integrates machine learning with a web-based interface to enhance accessibility and usability. The frontend is built using HTML5 and Jinja templates, dynamically rendering the welcome screen, questionnaire form, and result page. The backend is developed using Flask, a lightweight Python web framework, which manages routing, form submissions, and session handling. The trained ML model is serialized using Pickle, allowing seamless loading for real-time predictions. Additionally, FPDF is employed to generate downloadable PDF reports for users, summarizing their input, risk score, and relevant recommendations. This system not only demonstrates the practical application of ML in healthcare but also offers a complete pipeline from data input to result delivery. Future work includes expanding the dataset, integrating additional health indicators, and deploying the solution for mobile use or real clinical environments.

Details

Logo image