Flight delay prediction method based on ARIMA model

Abstract

The invention relates to a flight delay prediction method based on an ARIMA model. The method comprises the following steps of stepS1, establishing a data set: flight data is collected and the flight data includes flight information, time information, airport information and delay time information; stepS2, analyzing a data set characteristic: differences of airports and airlines and an influence of weather on the flight are analyzed; stepS3, establishing the ARIMA model: establishing of the ARIMA model includes a mobile average process, an autoregressive process, an autoregressive mobile average process and an ARIMA process; stepS4, selecting an optimum ARIMA model: the established ARIMA model is verified and an optimum parameter is selected; stepS5, predicting flight delay: any flight data is acquired and then an optimum ARIMA function is selected to be a model delay function, a multiple-linear regression function is selected to be a weather delay function and a final prediction result is acquired after addition of the functions. By using the method, high prediction precision can be realized, usage time is shortened and flight delay can be effectively predicted.

Description

technical field [0001] The invention relates to the technical field of flight information analysis, in particular to a flight delay prediction method based on an ARIMA model. Background technique [0002] Effective data mining algorithms and analysis strategies can extract valuable information for companies or individuals by acquiring prior knowledge and help them make further decisions. Among those fields involving big data, flight delay prediction has attracted extensive attention. The importance of risk management for flight delays has become apparent in recent years. For example, the occurrence of the appalling MH370 flight accident, as well as the huge economic losses caused by domestic flight delays have brought great dissatisfaction to airlines and passengers, which has pushed flight risk management to a very urgent position. Therefore, no matter from the aspect of safety factor or economy, a more effective flight delay prediction model is very necessary. [0003] ...

AI Technical Summary

Problems solved by technology

[0003] However, given the characteristics of flight data (e.g., large data volume, diversity, etc.), it is difficult to predict flight delays with high accuracy while keeping computational complexity and delays within an acceptable range

In addition, among the characteristics that affect flight delays, factors such as weather may change dynamically

Images