Search Results
You are looking at 1 - 1 of 1 items for
- Author or Editor: J. van Heteren x
- Refine by Access: All Content x
Abstract
Most of the common statistical tests are based upon statistically independent data. Many observations of physical phenomena are statistically dependent, as is indicated by the sample-autocorrelation function. In this paper, a test is presented for observations that are not statistically independent. The test is a modified form of the classical χ2-procedure and is based upon a transformation of the observations to statistically independent data assuming a second-order autoregressive model for describing the autocorrelation of the original observations. The test is applied to wind waves on sea, i.e., for testing on normality and on a “Gram-Charlier distribution” of sea surface deflection. The results show that the instantaneous values of sea surface deflection during swell conditions satisfy the assumption of a normal distribution and those during sea conditions are approximately normal, though a better approximation is provided by the “Three-term Gram-Charlier A Series distribution.”
Abstract
Most of the common statistical tests are based upon statistically independent data. Many observations of physical phenomena are statistically dependent, as is indicated by the sample-autocorrelation function. In this paper, a test is presented for observations that are not statistically independent. The test is a modified form of the classical χ2-procedure and is based upon a transformation of the observations to statistically independent data assuming a second-order autoregressive model for describing the autocorrelation of the original observations. The test is applied to wind waves on sea, i.e., for testing on normality and on a “Gram-Charlier distribution” of sea surface deflection. The results show that the instantaneous values of sea surface deflection during swell conditions satisfy the assumption of a normal distribution and those during sea conditions are approximately normal, though a better approximation is provided by the “Three-term Gram-Charlier A Series distribution.”
