الملخص الإنجليزي
The overlapping or similarity coefficient is defined to be a measure of the
similarity between two probability distributions. This coefficient is applied in
various fields, such as signal processing, econometrics, ecology and particularly in
meteorology. The objective of this study is to construct an index of similarity of
wind speed distributions based on the overlapping coefficients and to apply this
index to wind speed data in the Sultanate of Oman. We will use both parametric
and nonparametric approaches to estimate the overlapping parameters of two
distributions. First, in the parametric approach, we will fit parametric distributions,
such as Burr type II, to the wind speed data and then we will estimate the
overlapping coefficients. Whereas, in the nonparametric approach, we will
estimate the overlapping coefficient by the Kernel Density Estimator (KDE) using
the plug-in method. The main results are obtained in this thesis as follow.
Although, the Burr distribution fits most of wind speed in stations, but it is
iii
difficult be adequate for all the station's data. So, KDE can be an alternative
method to fit the data. Furthermore, the most of the coefficients and are
close to each other. Also, it is noticed in all the cases the measures have this
relationship of As well as , it has shown that the coefficients of KDE
and Burr distributions and 95% Bootstraps Intervals are agreements. For the degree
of similarity between pair stations, the distribution of wind speed between Masirah
and Rustaq stations is a high different, because there are difference in geography
area between these two sites. The most of the pair stations are moderate similar.
