Research On The Difference Of Flight Price Changes Based On Statistical Analysis

With the continuous improvement of people's living standards,continuous development of science and technology,and continuous social progress,tourism has gradually become an indispensable part of people's lives.Traveling by plane is the first choice for many people.However,the flight price is fluctuating,and the fluctuating price is a very complicated issue.Many reasons can cause the fluctuations of the price.Such as the quality of flight sales and the time from the date of departure,etc.The price forecasting technology of flights has huge commercial value and development prospects.The issue studied in this paper is whether there are significant differences in the price fluctuations between different flights during the flight price forecast process.When predicting the price of a flight,different forecasting models may be established for different flights,and each predictive model requires complex modeling and training processes.If there is no significant difference between the flights,then we can use the same price forecasting model to predict the price.This can save a lot of costs.This paper describes the feasibility of establishing the same prediction model for different flights through the use of statistical knowledge.This paper uses the Wilcoxon rank sum test in statistics to compare flights on different routes,different flights on the same route,international flights and domestic flights,international flights and international flight.Verifying that there is a significant difference in price fluctuations between different flights.Through the Wilcoxon rank sum test on the real data obtained from companies,It is concluded that the difference between the price fluctuations of different flights between the same routes is very small,The same predictive model can be applied to the predictive model.For different flights between different routes,the difference in price fluctuations is very significant and the conclusion of the same predictive model cannot be applied.