Kernel density estimation for heavy-tailed distributions using the champernowne transformation
read more
Citations
Composite Lognormal-Pareto model with random threshold
Skewed bivariate models and nonparametric estimation for the CTE risk measure
Modeling Extreme Events in Time Series Prediction
A UK best-practice approach for extreme sea-level analysis along complex topographic coastlines
Parameter recovery for the Leaky Competing Accumulator model
References
Density estimation for statistics and data analysis
Continuous univariate distributions
Theory of point estimation
A reliable data-based bandwidth selection method for kernel density estimation
Related Papers (5)
Frequently Asked Questions (7)
Q2. What is the recent study of the nonparametric estimator?
Scaillet [9] has recently studied non-parametric estimators for probability density functions which have support on the non-negative real line using alternative kernels.
Q3. What is the reason for the bad performance of the CHL estimator?
The bad performance of the CHL estimator is due to the fact that the transformation functions in this case always starts at 0 when α >
Q4. Why does the KMCE estimator overestimate the tail?
The authors have also seen that the KMCE estimator overestimates the tail, which is because the transformation function has a heavy Pareto tail.
Q5. Who improved the shifted power transformation for highly skewed data?
Bolancé et al. [4] improved the shifted power transformation for highly skewed data by proposing an alternative parameter selection algorithm.
Q6. What is the mean and variance of the classical kernel density estimator?
The mean and variance of the classical kernel density estimator isE [̂g(y)] = g(y) + 1 2 b2µ2(K)g ′′(y) + o(b2), (6)V [̂g(y)] = 1 NbR(K)g(y) + o ( 1Nb) . (7)The transformation kernel density estimator can be expressed by the standard kernel density estimator:f̂ (x) = T ′(x)ĝ(T (x))implyingE [ f̂ (x) ] = T ′(x)E [̂g(T (x))] = T ′(x) ( g(T (x)) + 12 b2µ2(K)∂2g(T (x))∂(T (x))2 + o(b2)) .
Q7. What does the standard statistical methodology do not weigh differently in the evaluation of an estimator?
Standard statistical methodology, such as integrated error and likelihood, does not weigh small and big losses differently in the evaluation of an estimator.