Apple Stock Price Forecasting

Overview

Stacked LSTM in Keras API is used to train Apple Stock (Close) Data to analyse and forecast the future stocks in Apple company.

• Model trained on Apple Stock Data that had been collected from Tiingo from Date 2016-07-28 till 2021-07-26 (or) download the same dataset used in this project.
• Trained Model (saved models/Stacked LSTM for Apple Stocks.h5) takes previous 100 days stock (Close) data as input and can forecast on (Close) stock on specified N number of days.
• Model trained on 717 records and tested on 340 records.
• Model trained on 15 epochs with early stopping and model check point on decreasing validation loss, there by Trian_loss: 8.7085e-04 - Val_loss: 0.0366.

Installation

The Code is written in Python 3.8.5. If you don't have Python installed you can find it here. If you are using a lower version of Python you can upgrade using the pip package, ensuring you have the latest version of pip.

• to create a virtual environment if you wish.

Activate virtual environment if created and to install the required packages and libraries, run this command in the project directory after cloning the repository:

pip install -r requirements.txt

All the Code Explanation of this Project had been explained here right from Data Collection till Forecasting.

Notebook Summary

• Collect Stock Data
• Load Data
• Train Test Split
• Data Sequencing
• MinMax Scaling
• Reshape into 3D data
• Model Creation
• Load Saved Model
• Analyse RMSE
• Plotting Results
• Forecast on future 30 days