| With the scale of airport transportation expanding, the level of airport noise pollution is worsening,therefore, installing airport noise monitoring system has been an important method for a lot of airportto monitor the surrounding noise environments. However, this kind of fixed monitoring point has highcosts, high environmental requirements and poor stability. Furthermore, airport noise monitoring datais interspersed environmental noise generated by other noise sources (such as wind noise, constructionnoise, etc.), therefore, assessing the reliability of monitoring data becomes particularly important.Based on the analysis of the characteristics of airport noise monitoring data, this paper presents anassessment model of airport noise monitoring data reliability based on evidence theory. The modelmakes use of methods of data mining to generate association rules among noise monitoring point data,and then makes use of the association rules to obtain basic probability assignment function, finallycombines these evidence by using Dempster combination rule and makes decision.Because of the defects of Dempster combination rule in combining conflict evidence, this paperproposes to use pignistic probability function to measure the degree of conflict between the evidencebased on the analysis of the measuring metrics about conflict, and then constructs a method of falseevidence recognition and combining conflict evidence. These methods are applied to the assessmentmodel of airport noise monitoring data reliability, so that with conflicting evidence the model can notonly identify the conflict of evidence, but also make use of the improved combination method to getthe correct fusion result.Experiments show methods of false evidence recognition and conflict evidence combinationproposed in this paper used at the airport noise monitoring data can effectively get better results, andthe improved assessment model has better accuracy and practicality used in real airport noisemonitoring data. |
PDF Full Text Request