A method and system for classifying and predicting aircraft slots based on two-way LSTM

Abstract

The invention discloses a two-way LSTM-based aircraft seat classification and prediction method, which belongs to the technical field of artificial intelligence. The method of the present invention first collects the flight information of the terminal building, adds tags to the flights according to the flight stops, and divides the flight information into different sequences according to the date; then preprocesses the flight information, and divides the flight information into training sets, verification set and test set; then use the flight information to build a bidirectional LSTM network model, use the training set to train the bidirectional LSTM network, test the model through the verification set, and adjust the network super parameter, so that the correct rate of the two-way LSTM network model is greater than the set threshold; finally, the real-time flight information is input into the trained two-way LSTM network model, and the predicted flight stop position is output. The invention also realizes a two-way LSTM-based aircraft stand classification and prediction system. The technical scheme of the invention can predict the seat category of the flight more accurately.

Description

Technical field [0001] The present invention belongs to the field of artificial intelligence, and more particularly to aircraft based on bidirectional LSTM based on aircraft unit classification prediction method and system. Background technique [0002] The position of the unit allocation is a complete problem with NP (Non-Deterministic Polynomial). Many airports in China still use manual allocation methods. However, with the continuous expansion of the airport scale, the unit distribution plan has grown index, which brings staff The huge workload, the intelligence and convenience of achieving the dispensing of the unit allocation is imminent. Under the premise of collecting a large number of historical data, the inherent law of data using machine learning technology is used to provide new ideas to resolve the installation problem. [0003] Machine assignment usually uses a variety of modern numerical solution algorithms such as traditional genetic algorithms, analog annealing al...

AI Technical Summary

Problems solved by technology

[0003] Various modern numerical algorithms such as traditional genetic algorithm, simulated annealing algorithm, etc. can be used for seat allocation, but they do not make effective use of historical data, and often have the disadvantages of slow convergence speed and difficulty in initial population generation.

Images