Flight Delay Prediction Using Machine Learning Models: A Case Study from the United States Airlines

The airline or aviation business is, without question, one of the fastest-growing ones right now. As the world's economy became more dependent on each other, this made flying one of the most regulated fields. There are a lot of financial and reputational losses for airlines when planes are late. However, the losses do not just affect the airlines; most of the losses are also felt by passengers. So, it is very important to be able to predict flights that will be late so that both customers and companies do not save money. The goal of this study is to use Naïve Bayes (NB), Decision Tree (DT), and Logistic Regression (LR) methods to build machine learning models that can look at and group flight delays. Our choice of features was based on categories, and it was Chi-Squared. The results showed that, all three models did well with the given information, with an average accuracy of about ۸۴%. The model that uses the LR model to predict the state of a flight's arrival does the best (۸۵.۱۴%). This study will help airlines, the government, and passengers by giving them accurate information about flights. This will cut down on economic losses and boost passenger trust.

کلیدواژه ها:

Flight Delay ، Machine Learning ، Naïve Bayes ، NB ، Decision Tree ، DT ، Logistic Regression ، LR