Spaces:
Sleeping
Sleeping
Create info.txt
Browse files
info.txt
ADDED
|
@@ -0,0 +1,44 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# MLP Neural Network LDL Regressor
|
| 2 |
+
|
| 3 |
+
This is a machine learning-based **LDL cholesterol estimation tool** developed using a **Multi-Layer Perceptron (MLP) regressor**.
|
| 4 |
+
|
| 5 |
+
## 🔍 How it works
|
| 6 |
+
|
| 7 |
+
This model estimates **Low-Density Lipoprotein (LDL)** cholesterol based on three standard lipid profile inputs:
|
| 8 |
+
|
| 9 |
+
- **Total Cholesterol (TC)** in mg/dL
|
| 10 |
+
- **Triglycerides (TG)** in mg/dL
|
| 11 |
+
- **HDL Cholesterol (HDL)** in mg/dL
|
| 12 |
+
|
| 13 |
+
From these, it internally calculates:
|
| 14 |
+
|
| 15 |
+
- **Non-HDL Cholesterol (NHC)** = TC - HDL
|
| 16 |
+
- **TG/NHC Ratio** = TG / (TC - HDL)
|
| 17 |
+
|
| 18 |
+
The model was trained on curated clinical lipid datasets and outperformed conventional LDL estimation formulas and other machine learning models in terms of adjusted R² and RMSE.
|
| 19 |
+
|
| 20 |
+
## ⚙️ Model Details
|
| 21 |
+
|
| 22 |
+
- Type: Multi-Layer Perceptron (Neural Network)
|
| 23 |
+
- Features used: TC, NHC, TG/NHC
|
| 24 |
+
- Trained using: `scikit-learn` MLPRegressor
|
| 25 |
+
|
| 26 |
+
## 🧪 Use Cases
|
| 27 |
+
|
| 28 |
+
- Research labs
|
| 29 |
+
- Clinical simulations
|
| 30 |
+
- Biochemistry education
|
| 31 |
+
- Digital health platforms
|
| 32 |
+
|
| 33 |
+
---
|
| 34 |
+
|
| 35 |
+
### 📜 Copyright
|
| 36 |
+
|
| 37 |
+
© 2025 **Dr. Saurav Nayak**
|
| 38 |
+
Assistant Professor
|
| 39 |
+
Department of Biochemistry
|
| 40 |
+
IMS & SUM Hopsital Campus-2
|
| 41 |
+
|
| 42 |
+
The KALP.AI Inititaive
|
| 43 |
+
|
| 44 |
+
All rights reserved.
|