Forecast and Early Warning of Regional Bus Passenger Flow Based on Machine Learning

Table 1

Advantages, disadvantages, and applicability of common short-term prediction methods.

Prediction method	Advantages	Disadvantages	Applicable conditions
ARIMA model	The model is simple and has the ability to correct local data trend	It is difficult to fit nonlinear problems	Medium- and short-term forecast
Exponential smoothing method	Flexible and simple operation	Less variables are considered and the accuracy of smoothing number is low	Medium- and short-term forecast
Trend extrapolation	The operation is simple and the fitting effect is good	It is difficult to guarantee the accuracy due to less variables	Short-term forecast
Multivariate regressive method	Multiple factors can be considered	Large amount of calculation and high requirement for data	Medium- and short-term forecast
SVM	The model is simple and the results need not be modified	It is difficult to consider the comprehensiveness of indicators	Medium- and short-term prediction of small samples
RF	It can process high-dimensional data without feature selection	The reliability of the attribute weight on the data is not high	Medium- and short-term prediction of small samples
LSTM	Strong nonlinear fitting ability	A large amount of data is needed for network training	Nonlinear prediction