Testing Serial Independence via Density-Based Measures of Divergence
read more
Citations
A Robust-Equitable Measure for Feature Ranking and Selection
Copula Correlation: An Equitable Dependence Measure and Extension of Pearson's Correlation
An Updated Literature Review of Distance Correlation and Its Applications to Time Series
The role of orthogonal polynomials in adjusting hyperpolic secant and logistic distributions to analyse financial asset returns
References
R: A language and environment for statistical computing.
Density estimation for statistics and data analysis
An Introduction to Copulas
Possible generalization of Boltzmann-Gibbs statistics
On a measure of lack of fit in time series models
Related Papers (5)
Frequently Asked Questions (12)
Q2. What are the future works mentioned in the paper "Testing serial independence via density-based measures of divergence autori" ?
From simulations, the “ Portmanteau ” approach to extend the results obtained on the single lags to more than one, reveals to be the best choice since it preserves size across lags.
Q3. What is the function invariant under strictly monotonic transformations?
The functional ∆ is invariant under strictly monotonic transformations if∆(f∗, g∗ · g∗) = ∆(f, g · g) for all h strictly monotonic.
Q4. What is the key assumption for the asymptotic theory concerning several serial independence tests?
The boundedness of the support of fl and g is a key assumption for the asymptotic theory concerning several serial independence tests such as those in Robinson (1991), Hong and White (2005) and Fernandes and Néri (2010).
Q5. What is the definition of the density estimator?
The density estimators described in the previous section can be employed in order to estimate the dependence functionals which allow to built serial independence tests.
Q6. What is the global performance of the Kullback-Leibler functional?
the integrated estimator of the Kullback-Leibler functional ∆1, combined with the Gaussian kernel density estimation techniques, provides the best global performance.
Q7. What is the significance of the uniform noise setting?
Since the results obtained under Gaussian and skew-t noise are quite similar, the uniform noise setting plays a very important role.
Q8. What is the rule to select r and s?
Regarding the rule proposed by Kallenberg (2009) to select κ and (rj , sj), let mn : + → + be a function of the sample size n defining the largest dimension for r and s.
Q9. What is the power of the test obtained with the copula-based density estimators?
Another interesting observation is that the test obtained with the copula-based density estimators seems to be more robust with respect to changes in the distribution of noise.
Q10. What is the way to extend the results on the single lags?
From simulations, the “Portmanteau” approach to extend the results obtained on the single lags to more than one, reveals to be the best choice since it preserves size across lags.
Q11. What is the main reason for the development of tests for serial independence?
This has motivated the development of tests for serial independence which are powerful against general types of dependence (omnibus tests; see Diks, 2009, pp. 6256–6257 for details).
Q12. What is the reason why a test performed on a single lag should have a?
In this case, thevariables Xt and Xt+l, for l > 1, are independent; consequently, a serial independence test performed on the single lag l should have a low power (approximately equal to α) if l > 1 and a high power if l = 1.