Modelling of Solar Radiation in Baleysa State, Nigeria Using the Kernel Regression Method

This study focused on modelling total solar radiation in Bayelsa State using the Kernel regression analysis. The objective of the study is to obtain an appropriate Kernel regression model for estimating total solar radiation in Bayelsa State and to determine whether time has a significant impact of total solar radiation. The source of data used for this study was the secondary data obtained from National Aeronautics and Space Administration (NASA). The data comprises of solar radiation values measured over a period of eleven years (2007-2017). The findings of the study found R-square value of 0.983 (98.3%), and bandwidth of 0.1970. The result of the Rsquare value indicates a strong adequacy of the model which implies that the model is useful for estimating total solar radiation. Also, this result implies that time (Month) was able to explain about 98.3% of total variation in total solar radiation. In addition, findings showed that time does not significantly impact on total solar radiation since the model obtained a p-value of 1.00 which falls on the acceptance region of the hypothesis assuming 95% confidence level.