Scaling Exponents of Time Series Data: A Machine Learning Approach

Abstract

In this study, we present a novel approach to estimating the Hurst exponent of time series data using a variety of machine learning algorithms. The Hurst exponent is a crucial parameter in characterizing long-range dependence in time series, and traditional methods such as Rescaled Range (R/S) analysis and Detrended Fluctuation Analysis (DFA) have been widely used for its estimation. However, these methods have certain limitations, which we sought to address by modifying the R/S approach to distinguish between fractional Lévy and fractional Brownian motion, and by demonstrating the inadequacy of DFA and similar methods for data that resembles fractional Lévy motion. This inspired us to utilize machine learning techniques to improve the estimation process. In an unprecedented step, we train various machine learning models, including LightGBM, MLP, and AdaBoost, on synthetic data generated from random walks, namely fractional Brownian motion and fractional Lévy motion, where the ground truth Hurst exponent is known. This means that we can initialize and create these stochastic processes with a scaling Hurst/scaling exponent, which is then used as the ground truth for training. Furthermore, we perform the continuous estimation of the scaling exponent directly from the time series, without resorting to the calculation of the power spectrum or other sophisticated preprocessing steps, as done in past approaches. Our experiments reveal that the machine learning-based estimators outperform traditional R/S analysis and DFA methods in estimating the Hurst exponent, particularly for data akin to fractional Lévy motion. Validating our approach on real-world financial data, we observe a divergence between the estimated Hurst/scaling exponents and results reported in the literature. Nevertheless, the confirmation provided by known ground truths reinforces the superiority of our approach in terms of accuracy. This work highlights the potential of machine learning algorithms for accurately estimating the Hurst exponent, paving new paths for time series analysis. By marrying traditional finance methods with the capabilities of machine learning, our study provides a novel contribution towards the future of time series data analysis.

Stochastic Patterns of Bitcoin Volatility: Evidence across Measures

“…The employed methodological approach however suffers inefficiencies associated to the limitations of each one of the employed tools. The limitations of the R/S analysis method involve the method's (i) sensitivity to data with non-stationary behaviors [25], (ii) inefficiency to provide reliable estimations of the Hurst exponent, for small size datasets [26], (iii) assumption of a fractal data structure [27], and (iv) sensitivity to changes in the dataset's sequence [28]. Regarding the SMA model employed, its critical limitations are associated to the model's (i) lag behind current market conditions, which may in turn lead to delayed signals for market entry or exit [29], (ii) inability to quickly adapt to rapid price changes [30], (iii) assignment of equal weights to all past data, thus not fully facilitating the significance of the information of the most previous data, and justifying why SMA may be less effective in the case of choppy markets [31].…”