| Abstrakt: |
Flight delays remain widespread problem in the airline sector. Delays may arise from various circumstances, including meteorological conditions, technical failures, air traffic control limitations, and airline-specific complications like crew scheduling concerns, disrupting passengers' itineraries, airline operations, and overall financial outcomes. Therefore, creating predictive models and operational techniques that assist airlines in optimising scheduling, and enhancing decision-making and consequences of delays is the solution to the problem. Although various predictive algorithms and studies have done to anticipate aircraft delays. However, many current models either insufficiently integrate extensive meteorological or concentrate exclusively on particular routes or regions which lacks the ability to generalise their conclusions to a wider context. This creates a deficiency to forecast delays on a broader more dynamic scale as weather patterns particularly severe and unforeseen conditions. Therefore, the work focus on the development of a cutting-edge system designed to focus on critical influence of meteorological conditions with an innovative methodology for predicting flight delays using modern machine learning (ML) and deep learning (DL) procedures. Here, ML algorithms like decision tree, logistic regression, and random forest along with proposed DL method called long short-term memory (LSTM) is applied on flight delay dataset. This paper aims to find and improve the substantially operational efficiency in airline sector by allowing airlines to predict and alleviate flight delays more precisely to hence enhancing passenger happiness and minimising financial losses. DL techniques facilitate identification of intricate, non-linear correlations within the data, whereas ML approaches enhance the optimisation and interpretation of the forecasting process. |
Full text from SpringerLink