| Intelligent traffic control system is one of the effective means to alleviate urban traffic pressure.The forecast of traffic flow is the basis of intelligent traffic control system,and its prediction accuracy directly affects the control effect of intelligent transportation system on road traffic.It is of great significance to study the traffic flow forecast of urban road.According to the historical data of road inspection bayonet,the traffic flow is forecasted.The main work includes:1.Traffic flow data are analyzed and processed.Traffic flow refers to the number of vehicles passing through the road inspection point within a certain period of time,with obvious periodicity,relevance and randomness.In the past,the univariate statistical prediction model often assumed that the traffic flow data obeys the Gaussian distribution,which may cause negative values when the predicted value of the traffic flow data is small.In order to solve this problem,this paper proposes the use of Gamma distribution.At the same time,considering the existence of fault data in the traffic flow data,the fixed threshold method is used to detect the data and then the fault data is processed.2.The GAMLSS univariate forecasting model based on Gamma distribution is proposed.The GAMLSS model is a generalized regression model.Considering the form and characteristics of the Gamma distribution,the Box-Cox transform method is used to transform the traffic data of the integer type into the real type data of the Gamma distribution and take it into the GAMLSS model.GAMLSS prediction model with Gamma distribution is used to select the model based on BIC criterion,and the unknown parameters of the model are estimated by the method of maximizing the likelihood function.The method of this paper is tested by using the traffic flow data of Dongbeilianhe Road in Dalian,and the validity of the method is verified.The results show that the method proposed in this paper has better effect on the prediction of traffic flow data. |
PDF Full Text Request