Determining the number of factors in a forecast model by a random matrix test: cryptocurrencies

Abstract

We determine the number of statistically significant factors in a forecast model using a random matrices test. The applied forecast model is of the type of Reduced Rank Regression (RRR), in particular, we chose a flavor which can be seen as the Canonical Correlation Analysis (CCA). As empirical data, we use cryptocurrencies at hour frequency, where the variable selection was made by a criterion from information theory. The results are consistent with the usual visual inspection, with the advantage that the subjective element is avoided. Furthermore, the computational cost is minimal compared to the cross-validation approach.

References in corpus (20)

• Modularity and community structure in networks
• Measuring Information Transfer
• Random Matrix Theories in Quantum Physics: Common Concepts
• Level-Spacing Distributions and the Airy Kernel
• Noise Dressing of Financial Correlation Matrices
• Granger causality and transfer entropy are equivalent for Gaussian variables
• Community structure in directed networks
• Universal and non-universal properties of cross-correlations in financial time series
• JIDT: An information-theoretic toolkit for studying the dynamics of complex systems
• Nonlinear time-series analysis revisited
• Anticipating cryptocurrency prices using machine learning
• Transfer Entropy as a Log-likelihood Ratio
• Structure and causality relations in a global network of financial companies
• Multivariate analysis and Jacobi ensembles: largest eigenvalue, Tracy--Widom limits and rates of convergence
• Scaling properties of extreme price fluctuations in Bitcoin markets
• On the top eigenvalue of heavy-tailed random matrices
• Factor modeling for high-dimensional time series: Inference for the number of factors
• Reducing the Bias of Causality Measures
• Approximate null distribution of the largest root in multivariate analysis
• Free Random Levy Variables and Financial Probabilities