A Correction Model For Airport Noise Prediction Based On Bayesian

Airport noise prediction plays an important role on the control of aircraft noise.The INM is representative of the traditional prediction model for airport noise prediction, based on the aircraft's NPD(noise-power-distance) curve obtained in the laboratory. It has a large error because it is lack of consideration of the effect of meteorological conditions on noise propagation, but its advantage is forecasting stable.Although the model based on machine learning considers the impact of the external environment, it relys on historical data and predicts instability. Therefore, it is meaningful to research the model for airport noise prediction of high precision.In order to solve the problem of airport noise prediction is not accurate and stable,this paper proposes a Naive Bayesian ensemble Model for airport noise prediction correction which learns the difference between the prediction value of traditional model and the actual monitoring value. The proposed model aims to revise the bias of traditional model caused by external factors. In addition, the training samples of the model is first clustered and these naive Bayesian models are built in multiple clusters in order to decrease the number of training samples and reduce the probability of strong dependence attribute's appearing. The experimental results show that this ensemble model for airport noise prediction has high prediction accuracy.But it has a poor robustness. To solve this problem, this paper proposes the tabu Bayesian network learning algorithm, which considers not only the attribute condition dependence, but also the expert knowledge constrains. The experimental results show that the proposed model has a high accuracy and robustness.In addition, in order to achieve the large number of prediction result,this paper studys the track vector algorithm and proposes airport noise batch prediction system,which contains the track vector algorithm, the track batch input function, NoiseMap,Bayesian model for airport noise prediction correction. The experimental results show that this system can predict the noise result accurately and quickly.