A Method for Predicting the Running Time of Bus Sections Based on Improved Deep Forest

Abstract

The invention provides a bus section running time prediction method based on improved depth forest. The method comprises the following steps: collecting bus route information, GPS information and section information, performing normalization processing for the collected information, inputting the processed data into the improved depth forest to perform training, wherein the improved depth forest is composed of two parts including convolution multi-granularity scanning and limited cascaded forest; selecting seven variables as eight-dimensional input vectors xi of a training sample (xi, yi), andselecting the current section running time as an output vector yi; selecting 70% of sample data as training sample and inputting the training sample into the improved depth forest, inputting the other 30% into an improved depth forest detection training result; taking MAPE as an evaluation index for the detection training result, wherein the smaller the MAPE is, the lower the prediction error is,and the more ideal the effect is. The improved depth forest can be more adaptive to data of different scales to perform training, moreover, can remedy a problem of high requirements on random accessmemory and computing facilities, and can predict more accurate bus running time.

Description

technical field [0001] The invention relates to the technical field of forecasting the running time of public transport vehicles, in particular to a method for predicting the running time of bus sections based on improved deep forests. Background technique [0002] With the rapid development of my country's economy and the acceleration of urbanization, the number of motor vehicles has increased rapidly, and traffic congestion, traffic emissions, and traffic accidents have intensified. Public transportation has become an effective way to solve the above problems with the characteristics of high capacity, low pollution, safety and speed. Vigorously developing public transportation plays an important role in improving the level of transportation services and changing travel modes. The prediction of bus running time is the key to induce residents to travel, improve travel efficiency and improve traffic service level. At the same time, the operation of public transport vehicles...

AI Technical Summary

Problems solved by technology

[0003] Considering that traditional deep learning (equivalent to deep neural network) requires a large amount of training data during training, and cannot be used for small-scale data tasks; and the route planning and running time of buses may be adjusted after a period of time

At the same time, the traditional deep neural network model is complex, there are too many hyperparameters, and it relies too much on the parameter tuning process

Images