A Chord Invariant Forecasting Method for Short-term Bus Passenger Flow

Abstract

A chord invariant prediction method for short-term bus passenger flow, comprising the following steps: firstly convert the chord invariant of short-term bus time series simulation; The prediction model SI-PFPM is derived, and then the genetic algorithm is used to optimize the assignment of each parameter in the SI-PFPM, and finally the prediction model is used to predict the future bus passenger flow sequence. The method of the invention does not require a large number of data samples, does not require large-scale training, has a small calculation amount, and is simple and easy to implement. Effective data processing can be carried out for actual data, and data can be used to realize calculation, training, prediction, and evaluation, and provide a short-term bus passenger flow prediction model with high accuracy and strong generalization ability.

Description

technical field [0001] The invention belongs to the cross technical field of intelligent transportation, computer and physics, and relates to a string invariant prediction method of short-term bus passenger flow. Background technique [0002] With the rapid development of artificial intelligence, intelligent algorithms are playing an increasingly important role in the research of passenger flow forecasting models. There are a variety of intelligent forecasting methods to predict bus passenger flow. Liu Lijuan et al. proposed a short-term passenger flow prediction method for BRT stations based on a deep neural network (DDN). The input features include time features such as the day of the week, the hour of the day, and whether it is a holiday or not. And scene features such as inbound and outbound, payment methods, as well as passenger flow characteristics such as historical average passenger flow and real-time passenger flow, and combine these features to train different stac...

AI Technical Summary

Problems solved by technology

Liu Lijuan et al. proposed a short-term passenger flow prediction method for BRT stations based on a deep neural network (DDN). The input features include time features such as the day of the week, the hour of the day, and whether it is a holiday or not. And scene features such as inbound and outbound, payment methods, as well as passenger flow characteristics such as historical average passenger flow and real-time passenger flow, and combine these features to train different stacked autoencoders (SAE for short) to further initialize DNN, and finally the hybrid model (SAE-DNN) for example analysis, this method provides a more accurate passenger flow prediction model for the passenger flow of four BRT stations in Xiamen, but for non-BRT stations, it is impossible to obtain full deep learning features and lack of data Difficult to predict passenger flow at bus stops

Li Wenquan et al. used the least squares support vector machine to establish a short-term bus passenger flow forecasting model for a certain line in Changchun. Considering the influence of passenger flow at upstream and downstream stations, historical passenger flow at the same period, and historical passenger flow on the prediction performance of the model, the example analysis results show that in After the upstream and downstream stations and the historical passenger flow in the same period are set in the multi-input variables, the prediction performance of the forecasting model is improved, but the shortcoming is that there are many factors affecting the short-term passenger flow of the bus station and it is difficult to analyze. This method only considers the factors that affect the short-term passenger flow few factors

The error of this model is small, considering the trend of passenger flow time series data itself, but the measurement of external factors affecting passenger flow is a bit rough, and it is only inferred from the change of fractal dimension

Images