A Novel Airfare Dataset To Predict Travel Agent Profits Based On Dynamic Pricing

Profit prediction in the aviation industry is an im-portant process to increase competitiveness in business. However, it is challenging because it relies heavily on good quality datasets. A good quality dataset can increase the linear correlation between features and can improve the coefficient of determination is the proportion of variation in the dependent variable that can be predicted from the independent variable. The purpose of this research is to create a novel flight ticket dataset where we add three new features to the general flight ticket dataset. We add NetPembayaran, SalesAgent and Profit features to our dataset. The approach used in making this dataset is to collect flight ticket sales transaction data, flight schedule lists, and flight ticket sales history data from various available sources, and integrate them into one comprehensive dataset. Our novel dataset successfully increases the linear correlation between features and can also increase the coefficient of determination. We also evaluated the model we used for profit prediction using RMSE.