Intention process mining using a context-aware Hidden Markov Model

Abstract

The emerging availability of digital devices that can be utilized for activity tracking and context sensing opened new opportunities for context awareness and user intention recognition. Mainly, the opportunity to use generated data during user operational process execution and understand the intention behind its behavior under a given context. The hidden Markov model (HMM) has been widely used in many fields, such as speech recognition and computational biology. It can be seen as a class of stochastic processes that have a finite-state structure. It then offers a good promise for their applications in process mining. Then, much research has been done to build generic process models based on the study of user behavior captured during the implementation of operational processes. But they merely considered the relationship between the observed activities and their sequences. They ignored the implicit intention and the surrounding context, conditioning the user’s behavior when triggering the actual process. Consequently, the objective of this research was twofold. First, we specify a context HMM for intention mining in an unsupervised manner. Secondly, we upgrade the resulting model within context awareness property. Finally, we evaluated these models in a case study with a travel activity dataset. The experiments revealed that intention mining within a context-aware model had better precision in discovering the correct intentions.

An Intensive Spectrum for Intention Mining Analysis

“…The basic approaches of machine learning are supervised machine learning and unsupervised machine learning. The use of probabilistic models to evaluate the most likely intentions behind traces of activities, specifically Hidden Markov Model (HMM) [60]. The authors contrasted user actions with the already-existing purposeful process model while concentrating on the supervised approach to determining the intentions underlying users' actions.…”