metadata
metadata:
language: en
library_name: scikit-learn
pipeline_tag: tabular-classification
inference: true
tags:
- diabetes-prediction
- healthcare
- classification
- sklearn
license: mit
author: Thyagharajan K K
flags:
- cpu
- allow_fork
- allow_metrics
emoji: 🩺
colorFrom: green
colorTo: pink
🩺 Diabetes Prediction Models
This directory contains machine learning models for diabetes prediction using structured clinical features.
The models are trained on standardized tabular data and evaluated using stratified K-fold cross-validation with robust statistical reporting.
## 📂 Directory Structure
diabetic_models/
│
├── README.md
├── CITATION.cff
├── LICENSE
├── logistic_reg_diabetic.pkl
│
└── SGD_scaler_diabetic_models/
├── README.md
├── SGD_model.pkl
└── SGD_scaler.pkl
🧠 Available Models
🔹 Baseline Model
- Standard Logistic Regression
- Batch optimization
- Used for benchmarking
🔹 Primary Model
- SGD-based Logistic Regression
- Supports:
- Epoch-wise validation
- Early stopping
- Adaptive learning-rate scheduling
- Serialized with its preprocessing pipeline
⚙️ Preprocessing
All models rely on feature standardization using StandardScaler.
The scaler is stored separately and must be applied during inference.
📜 License
This repository is released under the MIT License.
See the LICENSE file for full terms.
📖 Citation
If you use these models in academic work, please cite using the CITATION.cff file.
👤 Author
Dr. Thyagharajan K K
Professor & Dean (Research), RMD Engineering College
📧 kkthyagharajan@yahoo.com