# AI-Based Diamond Price Prediction and Classification

This project aims to predict diamond grading prices, GIA-certified prices, and classification-based changes in diamond attributes using machine learning models. The system processes diamond attributes from engineer plans and provides AI-driven insights into pricing and parameter variations.

---

## 🚀 Project Overview

The goal of this project is to develop an AI-driven pipeline for automating diamond grading and certification price predictions. The system leverages machine learning models to analyze historical diamond data, predict pricing estimates, and provide classification-based recommendations.

### 🔹 Features:
- Predict **GIA-certified prices** and **grading prices** for diamonds.
- Classify and recommend **potential changes in diamond parameters**.
- Analyze **historical data trends** for better forecasting.
- Provide **real-time AI predictions via a web-based interface**.

---

## 📌 Problem Statement

Diamond pricing and grading involve complex, time-consuming manual evaluations. This project automates the process by utilizing **machine learning models** to predict pricing, detect parameter changes, and generate valuable insights for decision-making.

---

## ⚙️ Tech Stack

| Component        | Tools & Technologies |
|-----------------|---------------------|
| **Data Collection** | Python (Requests, SQL) |
| **Data Preprocessing** | Pandas, NumPy, Scikit-learn Pipelines |
| **Model Development** | Scikit-learn, XGBoost, LightGBM, TensorFlow/PyTorch |
| **Model Evaluation** | Scikit-learn metrics (RMSE, MAE, R²), Evidently, Prometheus |
| **Deployment** | Flask, FastAPI, Docker |

---

## 🛠 Setup & Installation

### 1️⃣ Create a Virtual Environment
```bash
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate
```

### 2️⃣ Install Dependencies
```bash
pip install -r requirements.txt
```

### 3️⃣ Run the Application
```bash
python app.py
```

OR (if using Docker)
```bash
docker-compose up --build
```

---

## 📊 Application Workflow

### **🔹 Prediction Module**
**Input:** Diamond parameters from engineer plans:  
`Tag, EngCts, EngShp, EngQua, EngCol, EngCut, EngPol, EngSym, EngFlo, EngNts, EngMikly, EngLab, EngAmt`

**Process:**
1. **Historical Learning:** AI model learns from past diamond data.
2. **Training:** Identifies patterns linking diamond attributes to final pricing.
3. **Deployment:** Predicts `GrdAmt, ByGrdAmt, GiaAmt` for new inputs.

**Output:**
- AI-generated price estimates with **>95% accuracy**.

---

### **🔹 Classification Module**
**Input:** Engineer Plan data with additional attributes:  
`Tag, EngCts, EngShp, EngQua, EngCol, EngCut, EngPol, EngSym, EngFlo, EngNts, EngMikly, EngLab, EngAmt, Carat, Black_Code, White_Code`

**Process:**
1. **AI Model Training:** Learns from past cases where Carat, Black_Code, or White_Code led to different outcomes.
2. **Alert Generation:** Detects discrepancies in new inputs and suggests corrections.

**Output:**
- **Alerts and recommendations** for potential adjustments in diamond parameters.

---

## 📂 App Structure

```
.
├── app.py                    # Flask application
├── templates/
│   ├── index.html            # Home page
│   ├── output.html           # Prediction result display
├── uploads/                  # Uploaded diamond datasets
├── Model/                    # Trained AI models (joblib)
├── Label_encoders/           # Encoders for categorical variables
├── requirements.txt          # Dependencies
├── Dockerfile                # Containerization setup
├── README.md                 # Documentation
```

---

## 🌟 Key Features

✔ **GIA Price Prediction** – Estimates diamond grading costs.  
✔ **Parameter Classification** – Identifies changes in carat, shape, and other factors.  
✔ **Real-time AI Predictions** – Instant price estimates based on historical data.  
✔ **User-friendly Web Interface** – Upload diamond data and get instant insights.  
✔ **Downloadable Reports** – Export predictions and analysis as CSV files.  

---

## 🛠 API Endpoints

| Endpoint | Method | Description |
|----------|--------|-------------