yazodi's picture
Upload 6 files
86ce230 verified
---
title: "๐Ÿ’Ž Gemstone Price Regression"
emoji: ๐Ÿ’ฐ
colorFrom: indigo
colorTo: blue
sdk: streamlit
app_file: app.py
pinned: true
license: mit
tags:
- regression
- machine-learning
- streamlit
- diamonds
- synthetic-data
---
# ๐Ÿ’Ž Gemstone Price Prediction App
This Streamlit app predicts the price of a gemstone using its physical and quality-related features.
## ๐Ÿง  Project Overview
- This project simulates a **gemstone pricing system** using synthetic tabular data.
- Features include: `carat`, `depth`, `table`, `x`, `y`, `z`, `clarity_score`, `color_score`, and `cut_score`.
- The target variable is **price** (USD).
- Model: **RandomForestRegressor**
- Trained on 1000 synthetic samples.
---
## ๐Ÿ“Š Performance
- RMSE: **605.16**
- Rยฒ Score: **0.9549**
---
## ๐Ÿš€ How to Run Locally
```bash
pip install -r requirements.txt
streamlit run app.py
๐Ÿ”ฎ Future Work
Area Improvement
Model Try XGBoost, LightGBM
Feature Engineering Interaction terms, log/carat scaling
Deployment Add API endpoint with FastAPI
Real-world Data Integrate real gemstone datasets
๐Ÿ“ Files
app.py: Streamlit interface
rf_model.pkl: Trained model
model_columns.pkl: List of input features
requirements.txt: Required libraries