LightGBM algorithm-based traffic forecast method

Abstract

The invention discloses a LightGBT algorithm-based traffic forecast method. The LightGBT algorithm-based traffic forecast method comprises the following steps of S1, acquiring traffic data, performingdata normalization, and dividing the data into training data and test data; S2, performing model training on the training data by a LightGBT algorithm, and determining a model parameter; S3, inputting a LightGBM model parameter and the test data, and forecasting traffic flow; and S4, performing error estimation on a LightGBM model forecast result, reducing forecast data, and outputting the forecast data. With the LightGBM model, the forecast time is substantially reduced as well as the forecast accuracy is improved, and the LightGBT algorithm-based traffic forecast method has better forecastperformance and generalization capability in forecast of expressway traffic.

Description

technical field [0001] The invention relates to technical fields such as machine learning methods and traffic volume forecasting, and in particular to a traffic volume forecasting method based on the LightGBM algorithm. Background technique [0002] Traffic volume forecasting is one of the core issues in the field of traffic management. Traffic volume is an important indicator of expressway traffic status. Short-term traffic volume forecasting is an important content of expressway management. The freeway traffic system is a complex system with multi-factor effects and multi-level structure. Due to the lack of cumulative data on freeway traffic volume, the traffic system is affected by multiple factors, and the mechanism of action is fuzzy. During the gradual formation of the freeway network, the state of the system It is difficult to accurately describe the structure, boundary, etc., which makes the expressway traffic volume have significant randomness and uncertainty in dyn...

AI Technical Summary

Problems solved by technology

[0004] In view of the above and one of the problems, the present invention provides a traffic volume forecasting method based on the LightGBM algorithm, which aims to solve the problem of difficulty in parameter optimization, large consumption of computing resources, and poor forecasting performance caused by composite model forecasting in traffic volume forecasting. poor problem

Images