Research On Delay Prediction Of High Speed Railway Train Based On Data Analysis

High-speed railways have developed rapidly in recent years and have become the first choice of inter-city transportation tools for the advantages of safety,speed,punctuality and low consumption.With the improvement on people's quality of life,passengers have put forward higher requirements for the service quality of high-speed rail transportation.Therefore,there is a higher requirement for the accuracy of the high-speed railway train operation and the prediction accuracy of the train delay time.The train daly prediction can assist the dispatcher to analyze and judge the daly situation of the train operation and the degree of daly propagation,and estimate the daly risk,as the basis for scheduling adjustment strategy formulation and transportation organization optimization.Therefore,it is of great significance to study the prediction model of the impact time of high-speed trains of the daly arrival point,which can provide favorable theoretical support for the driving command automation of highspeed railways.Firstly,it analyzes the background and significance of the topic,and summarizes the research status of the train's daly time for home and abroad.According to the cause of the train,the daly time of the train is divided into the initial daly time and the non-initial daly time.The train running process,the train daly propagation law,the train operation adjustment principle and measures are analyzed.Based on the train operation performance data,the train arrival time deviation quantity distribution,the initial daly times distribution,the non-initial daly time distribution and the influence of the train arrival time deviation on the train operation diagram structure are analyzed.According to the relationship between the arrival time of the train,the train operation sequence adjustment judgment model and the independent variable of the train late time prediction model are selected,and the pre-selection model is determined according to the type and distribution of the independent variables and the dependent variables.Comparing the model accuracy with the model prediction results,the best train operation and operation adjustment judgment model and the train late time prediction model are obtained.The specific practices are as follows:1.The train delay time can be divided into initial late time and non-initial late time according to the cause.The initial late time is caused by the initial interference factor,and not the initial late time and the train running relationship and the train operation chart structure.Train late propagation is divided into horizontal late propagation and longitudinal late propagation according to direction,and the interference has cumulative effect.In order to reduce the influence of interference propagation on train operation,train adjustment scheduling strategy plays a vital role.Therefore,the principles and strategies for train operation adjustment are elaborated and analyzed.2.Statistical analysis of the distribution of the train arrival time deviation,and its impact on the train operation map structure,that is,the actual and graphical changes in the train operation map balance.Comparing and analyzing the distribution of train arrival time deviation,the train daly times is divided into initial daly time and non-initial daly time according to the mathematical classification method,and the statistical analysis distribution.3.The actual train running sequence is the same as the illustrated train running sequence,that is,the train late running sequence adjustment factor is used as the model determining factor,and the train operation ordered adjustment judgment model based on the naive Bayes classifier is constructed.After comparing the predicted values with the experimental values,the model can achieve good prediction results,that is,application and research value.Since the train operation adjustment is an important factor affecting the degree and extent of the train's late time propagation,the train operation sequence adjustment judgment result is taken as one of the indicators of the train operation adjustment model independent variable,and the accuracy of the train late time prediction model is improved.4.The generalized linear regression model is used as the pre-selected model,and the train late-time prediction model is constructed according to the train operation relationship and the train operation adjustment judgment result.The data set is divided into a training set and a test set,and a 5-fold cross-validation method is adopted to train the model through the training set to test the accuracy of the set test model.Finally,the train late prediction model based on Logistic regression has a high degree of coincidence in the accuracy,size,trend and range from train late time prediction,and the prediction effected are better.This paper present and constructs a prediction method of train delay time based on train performance data and train-to-train operation relationship,which can reach a higher prediction accuracy within a certain allowable error range.