model_card.yaml

---
model_name: "Stock Price Predictor (XGBoost)"
version: "1.0"
model_type: "XGBoost Regressor"
description: |
  This model predicts future stock prices based on historical data.
  It is trained on Google stock price data (2020-2025) and forecasts 
  the Adjusted Closing Price (`Adj Close`).

  The model uses various stock indicators (Open, Close, High, Low, Volume)
  and time-based features (Year, Day, Month, Weekday) with cyclic encoding.

  **Ideal Use Case**: Time-series forecasting for stock prices.

license: "MIT"
dataset: 
  name: "Google Stock Price Data (2020-2025)"
  source: "Kaggle - https://www.kaggle.com/datasets/mzohaibzeeshan/google-stock-price-data-2020-2025-googl"
  
architecture: 
  framework: "XGBoost"
  encoding: "Sin-Cos Encoding for Month/Weekday, One-Hot Encoding for categorical data"

metrics:
  MAE: 0.212
  RMSE: 0.357
  R2_Score: 0.9998

deployment:
  huggingface: "https://huggingface.co/spaces/your_username/google-stock-predictor"
  github: "https://github.com/your_username/google-stock-predictor"
  api: "FastAPI/Gradio"

usage:
  installation: |
    ```python
    # Install dependencies
    !pip install xgboost pandas numpy joblib transformers torch
    ```

  use_pipeline: |
    ```python
    from transformers import pipeline
    import joblib
    import pandas as pd
    import numpy as np

    # Load the model
    model = joblib.load("xgboost_stock_model.pkl")

    # Define a function to use it as a pipeline
    def predict_stock(features):
        data = pd.DataFrame([features])
        prediction = model.predict(data)[0]
        return f"📈 Predicted Adj Close Price: {prediction:.2f}"

    # Example input
    messages = {
        "Close": 200.5, "High": 201.2, "Low": 199.8, "Open": 200.0, "Volume": 1500000,
        "Year": 2025, "Day": 1, "Weekday_sin": 0.0, "Weekday_cos": 1.0, "Month_sin": 0.5, "Month_cos": 0.866
    }

    prediction = predict_stock(messages)
    print(prediction)
    ```

  load_model_torch: |
    ```python
    import torch
    import joblib

    model = joblib.load("xgboost_stock_model.pkl")

    def predict_stock_price(features):
        features = torch.tensor(features, dtype=torch.float32)
        prediction = model.predict(features.numpy().reshape(1, -1))
        return prediction[0]

    sample_input = [200.5, 201.2, 199.8, 200.0, 1500000, 2025, 1, 0.0, 1.0, 0.5, 0.866]
    predicted_price = predict_stock_price(sample_input)
    print(f"📈 Predicted Stock Price: {predicted_price}")
    ```

---