Skip to content

AnLong98/SIAP_2022

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

27 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Introduction

Dogecoin, IDENA and Shiba inu price prediction using ARIMAX and SVR alghoritms. Data was sampled daily from CoinGecko, Yahoo Finance and Google trends.

ARIMAX model as fitted on 90% of sample data and tested on 10% for Doge and Shiba. IDENA dataset was split in 70:30 ratio. Relevant attributes were taken into account with 3, 7, 14, 21, 30 and 60 day lag to test predictive value. Improvements in model predictivity were made by analyzing correlation heatmaps and introducing average variable value for lag windows along with standard deviation as predictors for one day ahead. ARIMAX hyperparameters were tuned with Auto Arima stepwise algorithm, criterion for model choice being AIC. Models were evaluated with mean average percent error (MAPE).

More info can be found in my paper

Sample result for IDENA, one day ahead:

bonitulja

Technologies

  • Python
  • Jupyter notebook

Authors

Predrag Glavas and Nikola Mijonic

License

MIT

About

Cryptocurrency price prediction with Python ML

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published