The influence of stock market indexes (S&P500 and Dow Jones) on cryptocurrencies prices
Date Issued
2022
Author(s)
Angelovski, Gorast
Todorovska, Ana
Rusevski, Ivan
Marojevikj, Jovana
Spirovska, Eva
Peshov, Hristijan
Vodenska, Irena
Chitkushev, Lubomir
Abstract
In this paper we analyze openly available time
series data for the prices of 18 cryptocurrencies and 2 stock
market indexes (S&P500 and Dow Jones).
First, we calculate the correlation values between the cryptocurrencies and indexes datasets.
Then, we use a state of the art time series prediction library
(XGBoost) in order to make prediction models for the daily
prices of all the cryptocurrencies, using the stock market index
datasets as input features in the training model.
We calculate metrics for the difference between the actual prices
and the prices predicted using our models.
Finally, we show the feature importance score that our model
attributed to each prediction model, and compare the score
between the three input features (S&P500 dataset, Dow Jones
dataset, and the actual cryptocurrency dataset).
series data for the prices of 18 cryptocurrencies and 2 stock
market indexes (S&P500 and Dow Jones).
First, we calculate the correlation values between the cryptocurrencies and indexes datasets.
Then, we use a state of the art time series prediction library
(XGBoost) in order to make prediction models for the daily
prices of all the cryptocurrencies, using the stock market index
datasets as input features in the training model.
We calculate metrics for the difference between the actual prices
and the prices predicted using our models.
Finally, we show the feature importance score that our model
attributed to each prediction model, and compare the score
between the three input features (S&P500 dataset, Dow Jones
dataset, and the actual cryptocurrency dataset).
Subjects
File(s)![Thumbnail Image]()
Loading...
Name
CIIT_2022_9.pdf
Size
895.29 KB
Format
Adobe PDF
Checksum
(MD5):f9b9bddc05b3eab6e22b69196c8cc5ac