A comparison of machine learning algorithms for diabetes prediction
importance 6/10
ai assisted
Artificial Intelligence in Healthcare diabetes care stage: research
Problem
Diabetes has no permanent cure, so early detection from patient data is needed.
Why it matters
Early detection of diabetes can support timely intervention in a widespread chronic disease.
What the AI did
Seven machine learning algorithms and neural network variants were compared on the Pima Indian Diabetes dataset (768 patients, 9 attributes) to predict diabetes.
Results
Neural network with two hidden layers achieved 88.6% accuracy; LR and SVM also performed well
Impact
Demonstrates ML can reach 88.6% accuracy on a standard diabetes benchmark; no deployment evidence stated.
Record
| Venue | ICT Express |
|---|---|
| Published | 2021 |
| Work type | article |
| Authors | Jobeda Jamal Khanam, Simon Y. Foo |
| Citations | 490 (OpenAlex, at import time) |
| References | 15 |
| Model or method | Logistic Regression, SVM, and neural networks (best: two hidden layers) |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1016/j.icte.2021.02.004 |
| Open access | gold |
| Keywords | diabetes prediction, machine learning, neural networks, Pima dataset, classification |
Importance rating
6
significant — thousands to millions affected
Diabetes is a chronic disease with no permanent cure; early detection matters, but abstract gives no scale or impact figures.