Symbolic Prediction Model for Wind Speed Based on Spatial Temporal Real Dataset

Abstract

Weather prediction is extremely important in many fields such as power generation and planning, air and marine navigation, and other fields which require accurate knowledge of the weather in a specific location. Wind speed (WS) is one of the most challenging weather variables to predict. Therefore, there are many approaches to predict WS including traditional approaches and up to artificial intelligence (AI) based approaches. The target is to obtain the highest possible accuracy in prediction at minimal error. In this paper, WS will be predicted using a modest symbolic model which aims to predict WS at a specific location based on WS of the nearest weather station to that place. This model depends on the displacement vector between the weather station and the place at which the WS is to be predicted. In the proposed case study, a symbolic prediction model (SPM) was evaluated on different places at Cairo governorate based on different scenarios. Error has been calculated which proved that the accuracy of prediction using the proposed SPM is analogous to other sophisticated techniques in the literature. This model is not limited to predict WS but can be extended to any stochastic variable that rely on the displacement vector between two points, such as temperature and air pressure.