Nonparametric density estimation for data analysis has been widely used because the classical parametric one has several problems. Frequency analysis of hydrologic data was performed based on nonparametric density estimation. It is very important to s...
Nonparametric density estimation for data analysis has been widely used because the classical parametric one has several problems. Frequency analysis of hydrologic data was performed based on nonparametric density estimation. It is very important to select bandwidth h for Kernel density estimation. In this study, Silverman's rule of thumb, Cross-Validation, and Plug-In methods were used to estimate bandwidth h. Because those methods took much time to calculate bandwidth h according to sample size, WARPing method was applied for reducing calculation time. Finally, flood quantiles of Goan site were estimated according to each method.