Fuzzy based weight to mine frequent patterns from human interaction in meeting using directed acyclic graph

摘要

Meeting is a gathering of people to exchange information and plan joint activities for achieving a goal through verbal interactions. In a good meeting, participants' ideas are heard, decisions are made through discussions and activities are focused on desired results. The challenging part is to mine the most relevant interaction pattern from the meeting. Tree structures are not able to capture all kinds of interaction and also not able to distinguish between ranks among the participants. Thus, meetings can be modeled as weighted DAGs, from which weighted frequent interaction patterns can be discovered. However, the weights can vary based on a participant's age and confidence of an action like commenting on a proposal. As this may not capture all interaction patterns, the rank of a participant is needed to be considered. This paper proposes a technique to calculate Fuzzy based weight of the participants based on age, rank and strength of comment and the existing weighted DAG is used to mine weighted frequent interaction patterns.