EsferSami/DataSynthis_ML_JobTask

Apple Stock Price Forecasting

This repository contains models for forecasting Apple stock prices using ARIMA and LSTM.

Overview

This project provides pre-trained models for predicting Apple (AAPL) stock prices:

• ARIMA Model – Classical time series forecasting using ARIMA with Box-Cox transformation.
• LSTM Model – Deep learning based forecasting using a trained LSTM network with a scaler.

Both models use the last 3 months of stock data to generate a 7-day forecast.

Inference Instructions

You can perform inference in one of two ways:

1. Run the provided inference notebooks

   Each model folder contains a ready-to-run notebook along with the pre-trained model files:

   • ARIMA Model:
     Folder: Apple-Stock-Price-Forecasting-ARIMA-Model
     Notebook: inference.ipynb (includes loading the ARIMA model and Box-Cox transformer, downloading recent AAPL data, and generating a 7-day forecast)

   • LSTM Model:
     Folder: Apple-Stock-Price-Forecasting-LSTM-Model
     Notebook: inference.ipynb (includes loading the LSTM model and scaler, preparing the last 60 days of data, and generating a 7-day forecast)

2. Use the code from the notebooks directly in your Python environment

   Each notebook contains fully commented code showing how to:

   • Download recent stock data (yfinance)
   • Load the pre-trained model from Hugging Face Hub
   • Preprocess data (Box-Cox for ARIMA, scaling for LSTM)
   • Run 7-day predictions
   • Generate a results table with forecasted prices

Note: If you want to directly run inference without notebooks, you can copy the code from the inference.ipynb files in each model folder. The notebooks also include instructions for installing required packages and setting your Hugging Face token.

License

This project is licensed under the MIT License.