New Grey Prediction Modeling Technique And Its Application In Air Pollution Prediction

Grey system theory is a systematic science that studies the uncertain of small samples,poor information.It has attracted the attention and research of many scholars since its inception.Grey prediction modeling technique is one of the core contents of grey system theory,and it is also a link between grey theory and practical application.However,the existing grey prediction modeling methods still have some defects in the modeling mechanism and modeling methods and it is difficult to meet the needs of practical problems.Therefore,this paper focuses on the practical problems of grey system theory in the new information priority modeling method and small-sample oscillation sequence prediction,and applies the above research results to the study of the prediction of air pollutant emissions with grey uncertain characteristics.The main research contents and innovations of the article are as follows:Firstly,a grey rolling prediction model considering new information priority is constructed.The new information priority principle is applied to the first,middle and last stages of the grey forecasting modeling process,making full use of the advantages of the new information in the system and fundamentally reducing the modeling forecasting error.Before modeling,the average weakening buffer operator is introduced to reduce the impact of periodic disturbance factors;in the modeling process,new data information is added,and the rolling prediction mechanism is used to continuously update the new and old information.After modeling,the error checking method of traditional grey model is improved,and the error of the first point data obtained from each rolling prediction is taken as the simulation error of GRPM（1,1）model.The structure of the traditional GM（1,1）is expanded,and a new grey rolling prediction model GRPM（1,1）is constructed.The modeling mechanism and modeling steps of the new model are described in detail.Secondly,a grey interval forecasting model for oscillation sequence is constructed.Based on the definition of small sample oscillation sequence and the existing grey model of interval grey number forecasting,a noval grey prediction modeling method with the idea of"interval range"is proposed.By designing the upper and lower bound function of the original sequence,the interval expansion of the oscillation sequence that is difficult to fit is carried out,and then the grey number whitening method is used to convert it into"area sequence"and"position sequence"with the same amount of information.DGM（1,1）is built based on the expanded sequence to obtain the predicted value,which can reflect the area and position of the original oscillation sequence,and then the predicted range value can be calculated by inverse derivation.Thirdly,the prediction of air pollutant emission in China.The emission characteristics of sulfur dioxide in China are analyzed comprehensively.The novel grey rolling prediction model is used to forecast and analyze the trend of sulfur dioxide emission.The new model uses the new information priority principle to effectively solve the problem of insufficient information utilization due to system disturbance factors,which has higher modeling accuracy than the traditional grey prediction model.Furthermore,according to the monthly PM_2.5concentration data of Beijing,the proposed grey interval prediction model is used to predict and analyze its development trend,which realizes the reasonable interval range prediction of Beijing’s PM_2.5concentration for a period of time in the future.According to the experimental results and the current status of air pollution,relevant control countermeasures and suggestions are put forward to help ecological environmental protection.