Methods of Introducing Continuous Process Mining to Service Management for Mobile APPs

Abstract

Precise and continuous service management is needed in today's mobile application development, with consideration of Continuous Delivery. This paper presents a framework to introduce sustained process discovery to service management for mobile APPs. Leveraging service logs accumulated at runtime, this framework bridges the gap between traditional process mining techniques and mobile APPs, demonstrating a log conversion method to deduce the missing case identifiers and gathering multiple versions of the process model, providing rich information for user behavior analysis, auto scaling of computing resources, etc. A case study is conducted based on a Chinese m-health APP to show our approach can provide comprehensive supports in many application scenarios. Experimental results demonstrate that our framework successfully handles with the incompleteness of original service logs by correctly deducing missing case identifiers in most cases.