A0660
Title: Predicting the price movement of cryptocurrencies using linear law-based feature space transformations
Authors: Marcell Tamas Kurbucz - Wigner Research Centre for Physics | Corvinus University of Budapest (Hungary) [presenting]
Antal Jakovac - Wigner Research Centre for Physics (Hungary)
Peter Posfay - Wigner Research Centre for Physics (Hungary)
Abstract: The aim is to investigate the effect of a novel method called Linear Law-based feature space Transformations (LLT) on the accuracy of intraday cryptocurrency price movement prediction. To this, first, sample series of the 1-minute interval price data of Bitcoin, Ethereum, and Ripple between 1 January 2021 and 2 August 2022 are obtained from Bitstamp. Then, the training and test sets of samples are separated. After that, LLT identifies the governing patterns (laws) of each input sequence in the training set by applying time-delay embedding and spectral decomposition. Finally, the laws of the training set are used to transform the feature space of the test set. The transformed test set is classified by traditional machine learning algorithms, such as decision tree (DT), support vector machine (SVM), and k-nearest neighbor (KNN) with 10-fold cross-validation. The results emphasize the potential of the LLT method in terms of both accuracy and calculation time in price movement prediction.
