Modeling wind speed data using the generalized positive exponential family of distributions

In this paper, we explore a general class of probability distributions to model the wind speed measurements. This class is named as a generalized positive exponential family of distributions and various existing models like Weibull, Rayleigh, and gamma are members of this family. Several important statistical properties of this family of distributions are thoroughly examined. The maximum likelihood estimation technique is proposed to estimate the model parameters. The flexibility and utility of the proposed model are demonstrated through the recent datasets on wind speed measurements from the National Oceanic and Atmospheric Administration and Inland Wind Farm. A detailed comparative analysis based on different statistical criteria highlights that the proposed model outperforms its existing counterparts, and provides greater accuracy while dealing with the real datasets. Although we focus only on wind speed distribution, the proposed model may also be useful for analyzing other important and related measurements.