The Logistic Lasso and Ridge Regression Algorithm-Based Airline Passenger Satisfaction Prediction Model

Airline passenger satisfaction is important for airline operation service quality management. When airline companies carry out advertisement campaigns or plan a marketing strategy, the resources and budgets are not unlimited. Thus, an airline can only focus on improving a few factors that drive passenger satisfaction. To understand the key satisfies for the young and the old adults, respectively, we leverage five airline passenger satisfaction methods to identify the key factors that explain the airline service satisfaction of different passengers. In particular, we investigate and compare the ridge and the Lasso regularization in terms of the resulting model’s sparsity and computational efficiency. The top three important factors that influence the old’s satisfaction are departure and arrival time convenience, legroom service, and baggage handling. Our findings indicate that the young people place a higher value on entertainment, while the old adults place a higher value on usefulness and comfort. The Lasso is the most accurate model with the overall error of 9.65% to predict the young passenger’s satisfaction, while the Best Subset with BIC with the overall error of 10% is the best mode for the old adults. It’s suggested that airline companies could use the Lasso model for predicting the airline satisfaction of the young people, and use the best subset with BIC for predicting the airline satisfaction of the old adults. The study findings would help the airlines improving their state-of-the-art operations to have outstanding service.