Investigation of hidden patterns in Qmatic Customer Journeys in order to minimize the average waiting time

Abstract

A queue management system is a critic component in any sector of business. In fact, all of the queue management systems have the same goal which is to minimize the waiting time for the customers who are in queues. The aim of this study is to investigate the predictive variables which explain waiting time in bank virtual queues of Qmatic customer journeys. The Knowledge Discovery in Databases which is the standard data mining process, is employed as the methodology to reach the target of discovering the hidden patterns. As the data mining engine a GBM model is used for waiting time estimation. One of the most substantial measures in creating the regression model is selecting the optimal feature sets for predicting waiting time. According to the results, the best results obtained when learning rate equals to 0.1, max depth equals to 4, min sample leaves equals to 2 and max feature equals to 0.3. For the training processes, 10- Fold cross-validation is applied. The overall accuracy of the model is 71%.