<![CDATA[Automatic equipment for monitoring weather conditions such as the Automatic Weather Observing System (AWOS) is urgently needed by a meteorologist for the purposes of serving aviation weather services at airports. One of the most important information besides the weather for flight services is wind speed. This study integrates AWOS and linear regression models to predict wind speed parameters for the next 12 hours. These parameters are the lowest, average, and highest wind speed. The computational load required for building and training the proposed model system is determined by the duration the computer executes the model training commands and generates predictions. The wind speed hours ahead is assumed to be influenced by the condition of the previous weather parameters. Therefore, in this study, a scheme was tested using the length of historical data of different weather parameters to predict the wind speed parameters for the next 12 hours. The predictions generated are in summary form, i.e., the lowest speed, average speed and highest speed in that period. After testing it was found that the duration of the computer to train the model is 1.2 seconds and to generate predictions is 1.1 seconds. Meanwhile, the best scheme for generating predictions is linear regression with a predictor of 12 hours which produces an RMSE error of 0.63, 1.14, and 3.07 for the lowest wind speed, average wind, and highest wind respectively. These results indicate that the proposed model only requires a light computational load and can produce accurate predictions of wind speed parameters for the next 12 hours.]]>
Copyrights © 2022
