Airport Noise Prediction Model Based On Collaborative Filtering Algorithm

With the rapid development of China's civil aviation industry, the problem of airport noise pollution increases prominently, which seriously hampers the development of the airport. In order to monitor the influence of airport noise effectively, airport noise monitoring system is widely used in more and more airports.The research on noise monitoring system with predicting function and the construction of the real-time noise prediction model have important significance for intelligent monitoring of airport noise.Based on the analysis of the existing monitoring system architecture, this paper presents a new airport noise monitoring system model, which can respond in real time by predicting. The model can predict the noise impact by introducing the real-time noise prediction module based on collaborative filtering, and provides strong support for the intelligent management of airport noise monitoring system.The spatio-temporal relationship among the monitoring data of airport noise monitoring system is studied, and an airport noise prediction model which combines the correlation property of the aircraft noise and flight path is proposed in this paper.The model based on the neighbor model improves the traditional Slope One algorithm,and overcomes the defects of the traditional algorithm that result from irrelevant items.Experiments on historical monitoring data set of a hub domestic airport show that compared with the conventional model, the efficiency of the improved Slope One algorithm has been improved significantly, and the proposed model can also ensure the accuracy of the prediction,Finally, this paper analyzes the airport noise influence factors, and puts forward a new cubic factorization model. The model extracts the characteristic factor of the aircraft and monitoring points from the airport noise monitoring data, and predicts noise based on the degree of matching characteristic. Experiments show that the cubic factorization model has higher prediction accuracy and better generalization ability compared with the traditional models.