Add model card for Delivery Time Estimation Model
Browse files
README.md
ADDED
|
@@ -0,0 +1,84 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: sklearn
|
| 3 |
+
tags:
|
| 4 |
+
- logistics
|
| 5 |
+
- xgboost
|
| 6 |
+
- sklearn
|
| 7 |
+
- tabular-regression
|
| 8 |
+
- eta-prediction
|
| 9 |
+
- delivery-time
|
| 10 |
+
- fairrelay
|
| 11 |
+
datasets:
|
| 12 |
+
- Cainiao-AI/LaDe-D
|
| 13 |
+
- electricsheepafrica/africa-synth-retail-and-ecommerce-last-mile-delivery-data-nigeria
|
| 14 |
+
license: mit
|
| 15 |
+
---
|
| 16 |
+
|
| 17 |
+
# FairRelay — Delivery Time Estimation Model
|
| 18 |
+
|
| 19 |
+
Part of the **[FairRelay](https://github.com/MUTHUKUMARAN-K-1/FairRelay)** AI logistics platform.
|
| 20 |
+
|
| 21 |
+
## Model Description
|
| 22 |
+
|
| 23 |
+
Delivery Time Estimation Model
|
| 24 |
+
|
| 25 |
+
**Type**: XGBRegressor Pipeline
|
| 26 |
+
**Framework**: scikit-learn Pipeline + XGBoost
|
| 27 |
+
**Task**: Regression
|
| 28 |
+
|
| 29 |
+
## Performance
|
| 30 |
+
|
| 31 |
+
- **R²**: 0.0785
|
| 32 |
+
- **MAE**: 169.16 minutes
|
| 33 |
+
- **RMSE**: 199.50 minutes
|
| 34 |
+
- **CV R² (5-fold)**: -0.0300 ± 0.0507
|
| 35 |
+
|
| 36 |
+
|
| 37 |
+
## Input Features
|
| 38 |
+
|
| 39 |
+
| Feature | Importance |
|
| 40 |
+
|---------|-----------|
|
| 41 |
+
| `distance_km` | 0.0974 |
|
| 42 |
+
| `hour_of_day` | 0.2686 |
|
| 43 |
+
| `day_of_week` | 0.1451 |
|
| 44 |
+
| `is_weekend` | 0.0534 |
|
| 45 |
+
| `num_packages` | 0.1910 |
|
| 46 |
+
| `num_stops` | 0.2445 |
|
| 47 |
+
|
| 48 |
+
|
| 49 |
+
## Usage
|
| 50 |
+
|
| 51 |
+
```python
|
| 52 |
+
from skops import io as sio
|
| 53 |
+
from huggingface_hub import hf_hub_download
|
| 54 |
+
|
| 55 |
+
# Download and load
|
| 56 |
+
model_path = hf_hub_download(repo_id="muthuk1/fairrelay-delivery-time", filename="model.skops")
|
| 57 |
+
untrusted = sio.get_untrusted_types(file=model_path)
|
| 58 |
+
model = sio.load(model_path, trusted=untrusted)
|
| 59 |
+
|
| 60 |
+
# Predict
|
| 61 |
+
import numpy as np
|
| 62 |
+
features = np.array([[0.0, 0.0, 0.0, 0.0, 0.0, 0.0]])
|
| 63 |
+
prediction = model.predict(features)
|
| 64 |
+
```
|
| 65 |
+
|
| 66 |
+
## Training Data
|
| 67 |
+
|
| 68 |
+
- **Cainiao-AI/LaDe-D**: Real last-mile delivery data from Shanghai (KDD 2023)
|
| 69 |
+
- **Africa Synth Last-Mile**: Synthetic Nigerian delivery data
|
| 70 |
+
- **FairRelay Synthetic**: Physics-informed synthetic data calibrated to FairRelay's deterministic formulas
|
| 71 |
+
|
| 72 |
+
## Part of FairRelay
|
| 73 |
+
|
| 74 |
+
FairRelay is an AI-powered logistics platform for fair load consolidation and dispatch:
|
| 75 |
+
- 🚚 5-agent load consolidation pipeline (KMeans + OR-Tools CP-SAT)
|
| 76 |
+
- ⚖️ 8-agent fair dispatch pipeline (Gini optimization)
|
| 77 |
+
- 📊 XGBoost effort prediction + Thompson Sampling bandit
|
| 78 |
+
- 🌱 EV-aware routing with battery constraints
|
| 79 |
+
|
| 80 |
+
Built for **LogisticsNow Hackathon 2026** — Challenge #5: AI Load Consolidation
|
| 81 |
+
|
| 82 |
+
## License
|
| 83 |
+
|
| 84 |
+
MIT
|