/stockprice-analysis-and-predictions

Primary LanguageJupyter Notebook

Stockprice Analysis and Predictions

This tool is meant to gain data from the Yahoo finance page and display some historical features of a stock trend within a given period of time

Some Results of the Stockprice analysis and predictions:

  • The best model after testing over 10 different stocks is always the GradientBoosting Regressor
  • Even Averaged models are combined not better the the GradientBoosting Regressor alone
  • The accuracy depends strongly on the volatility of each stock (the more volatile the less better/accurate the prediction)
  • The range of the accuracies for different tested stocks like Microsoft, Apple, Audi and many others are in a range of 97 - 99.99 % for the best model (GradientBoosting Regressor)
  • other models performed from about 90 % for very volatile stocks to up to 99.6 % for very stable one

Local Installation

Clone the repo (or simply download it)

git clone https://github.com/JanMarcelKezmann/stockprice-analysis-and-predictions.git

Install requirements

(navigate into the new folder)
pip install -r requirements.txt

Run with JupyterNotebook or JupyterLab

Just open the .ipynb code in a Notebook of your choice and run it.

Example: Microsoft