Research on flight arrival delay prediction based on support vector machine

Based on support vector machine, this paper proposes a model for predicting flight arrival delays through flight plan by analyzing flight data. By selecting one day's domestic flight plan and flight operation data, it analyzes the data of flight plan arrival and departure time, flight actual arrival time, flight type (passenger flight or cargo flight), constructs and selects feature vectors, including the planned flight time, the planned arrival time period, the planned departure time period, the planned arrival time in day or night, the planned departure time in day or night, the planned arrival in the peak time period, the planned departure time in the peak time period, the passenger or cargo flight. And it establishes a prediction model. Through case study, it found that flight planning has an important impact on flight arrival delays. On the one hand, when limited and non-repeated random samples are selected for training, one constant value has a greater impact on the prediction accuracy. The constant value is negative, indicating that the actual arrival time of the flight is later than the planned arrival time, the more stable the forecast results are. On the other hand, when the amount of sample data is large, the smaller the constant value, the more accurate the prediction result. At the same time, the random arrangement of the sample data has less influence on the prediction result, and the prediction result is stable.