Flight delay prediction for commercial air transport: A deep learning approach

Hình ảnh có liên quan

The motivation for this study was originated from a collaboration research with PEK, aimed at developing a practical data analytics-driven method to provide an accurate real-time prediction of flight delays. High-dimensional data from PEK between Jan 2017 and Mar 2018 were analyzed to capture the key factors influencing flight delays.

Benefiting from the flight information system and large dataset, this study aims to employ a practical DBN-SVR method and identify a novel set of micro influential factors, enabling aviation authorities to explore the underlying behavior and mechanism of flight delays. Novel influential factors, such as air route situation and crowdedness degree of airports (based on the number of flights and passenger flows), were introduced and examined through a multifactor approach. The results showed that these novel factors are indeed of high importance for flight delay prediction accuracy.Our proposed multifactor approach can significantly enhance the overall service level of commercial airports and airlines by providing accurate real-time flight information that consequently reduces passenger anxiety and complaints.

Collected and summarized from the source below by Giang Tan  https://db.vista.gov.vn:2095/science/article/pii/S1366554518311979#b0060