ProLift: A Web Application to Discover Causal Treatment Rules From Business Process Event Logs

Abstract

Causal process mining is a sub-field of process mining belonging to a family of techniques related to the field of business process management(BPM). The main goal of causal process mining is to utilize real process execution logs and causal machine learning techniques in order to discover, analyze and improve business processes. In this Master’s Thesis, we provide a detailed overview of a web-based application and all of its components, developed by the thesis author. The main goal of the application is to utilize the latest discoveries in causal process mining techniques that are capable of discovering and quantifying cause-effect relations. The additional goal of the application is to provide end users with a responsive and user-friendly interface, which allows them to discover treatment rules from a business process event logs and to display these treatment rules in an easy-to-understand manner. The Web application outlined in the thesis implements an approach to discover causal treatment rules proposed by Bozorghi et al. This approach uses uplift trees to discover rules that relate a treatment (e.g. giving a phone call to a user) with an increased probability that a positive outcome will be achieved (e.g. the customer will be satisfied with the service they receive).