Air traffic delay prediction method of PR-LSTM algorithm combined with attention mechanism

Abstract

The invention discloses an air traffic delay prediction method of a PR-LSTM algorithm combined with an attention mechanism. The method comprises four modules, namely a PageRank algorithm module, a sequence-to-sequence LSTM (Long Short Term Memory) module, an attention mechanism module and a delay prediction result display and release module. According to the airport delay prediction method, time correlation and spatial correlation of air traffic are considered at the same time, the size of a spatial correlation coefficient is calculated by adopting a PageRank algorithm, multistep prediction of airport delay is realized by adopting a sequence-to-sequence LSTM model, and the prediction precision of the model in multistep prediction is ensured by adopting an attention mechanism.

Description

technical field [0001] The invention belongs to the technical field of intelligent transportation, and in particular relates to an air traffic delay prediction method combined with a PR-LSTM algorithm of an attention mechanism. Background technique [0002] In recent years, with the development of the economy and the further improvement of the technical level, people's travel methods have undergone major changes, and air travel has gradually become the choice of travel for ordinary people. In 2018 alone, civil aviation transportation companies in the industry completed 11.5352 million transport flight hours, an increase of 8.9% over the previous year; they completed 611.7377 million passenger trips, a year-on-year increase of 10.9%. [0003] The ever-increasing demand for travel has brought enormous pressure to airlines. In the case of existing airport resources, in order to meet the needs of passengers, airlines can only meet the needs of passengers by continuously increas...

AI Technical Summary

Problems solved by technology

[0008] However, the inventors found that the driving traffic flow prediction method is more focused on ground traffic, and air traffic flow is difficult to grasp as a whole due to its more complex network structure and dynamics.

Although there are methods for studying single flight delay prediction, the research on the data-driven flight delay prediction method considering the entire airport network still needs to be perfected.

Images