Construction of Advanced Machine Learning Models for Air Traffic

This book explores the application of various time series and machine learning techniques to model and forecast domestic airline traffic. It provides a comprehensive study of traditional and modern predictive approaches. It presents an extensive literature review on airline traffic modeling, covering traditional time series methods(Holt's Winter, ARIMA, SARIMA) alongside advanced machine learning techniques(FFNN, MLP, LSTM). A comparative analysis of these methods, highlighting their strengths and limitations, is also included. Further, it explores the Bayesian estimation of SARIMA model parameters. The estimated parameters and predictions are compared with the traditional maximum likelihood approach. It extends the research by introducing mixture models, hybrid approaches, and simple averaging techniques to enhance predictive accuracy. The effectiveness of these models is evaluated through comparative analysis.

Autorentext
Dr. Mounika Panjala, an M.Sc. (Applied Statistics) and Ph.D. (Statistics) graduate from Osmania University, is a faculty member at the University of Hyderabad, Hyderabad, Telangana, India. She has excelled academically, qualifying UGC-NET, GATE (Statistics), and TSSET, with research focused on "Data Modelling through Machine Learning".