Design and evaluation of a process-aware recommender system based on prescriptive analytics

Process-aware Recommender systems (PAR systems) are information systems that aim to monitor process executions, predict their outcome, and recommend effective interventions to reduce the risk of failure. While a PAR system is composed by monitoring, predictive analytics and prescriptive analytics, the lion's share of attention in the recent years has been on the first two, overlooking the last. It seems that process participants are tacitly assumed to take the 'right decision' for the most appropriate corrective actions in case of failure's risks. Unfortunately, the assumption of selecting an effective corrective action is not always met in reality. When selecting an intervention, this is mainly based on human judgment, which naturally relies on subjective process' perceptions, instead of objective facts. Experience has shown that, when a fact-based predictive analytics is followed by subjective prescriptive analytics, the positive effect of good predictions are nullffied by inconclusive corrective actions, yielding no final improvement. This paper discusses a PAR system that features a data-driven prescriptive analytics framework, which puts aside subjective options and focuses on factual data. The effectiveness of the proposed solution is assessed through the process of a reintegration company, showing a potential increase of customers that find a new job.