Technical Sharing Session #4

Customer Behavior Prediction: Random Forest Approach

Title	Customer Behavior Prediction: Random Forest Approach
Presenter	Dr Liu Changchun
Presenter	Research Fellow, Work Package 1
Date	4 March 2022, Friday
Sypnosis	The foundation of a good pricing strategy is to better understand customer choice behaviour. It is interesting to use the historical data to learn the behaviour of customers’ choices. If we can learn customer choice accurately, we can simulate the profits of various pricing policies and evaluate the performance of each policy. In this talk, we will mainly show how we use the random forest to predict which fare type the customer selects. In this prediction, we only use general information without consideration of personal information. We will also analyse the data set to find the reasons why the predictions of some instances are inaccurate.