Using Multiple Regression and Artificial Neural Network Approach for Modeling Airport Visibility

According to flight safety statistics, low airport visibility is one of the reasons that cause flight accidents. Low airport visibility can be originated from many climatic factors, including thick fog, rain or thick haze for on ground situation and midst of cloud The main objective of the study is to construct a model which has the capability for forecasting airport visibility by using the weather data collected from weather division located at southern Taiwan dating from 1984 to 2004, a total of 21 years. In order to reach the objective, correlation coefficient and F test are calculated for determining the possible influence factors to the airport visibility. The determined influence factors are then applied to multiple linear regression and back-propagation artificial neural network for modeling airport visibility. From the results of the influence factor determination, humidity, visual hazard, temperature, and atmospheric condition are statistically significant for the airport studied herein. By comparing the coefficients of determination, one can say that the back-propagation artificial neural network approach is superior to the conventional regression method. Finally, the results from both models are used for estimating class predictive accuracy, the accuracy are 92.6% and 69.3% for the airport visibility under 1,800 meters and 93.4% and 76% for the airport visibility between 1,801 meters and 3,600 meters respectively for backpropagation artificial neural network and multiple linear regression models built in this study.