An implementation of efficient techniques for tree based mining in human social dynamics

Abstract

Meetings are an important communication and coordination activity of teams: status is discussed, new decisions are made, alternatives are considered, details are explained, information is presented, and new ideas are generated. As such, meetings contain a large amount of rich project information that is often not formally documented. Capturing all of this informal meeting information has been a topic of research in several communities over the past decade. In this work, data mining techniques are used to detect and analyze the frequent interaction patterns to discover various types of knowledge on human interactions. An interaction tree based pattern mining algorithms was proposed to analyze tree structures and extract interaction flow patterns for meetings. The work extends for tree based mining algorithm proposed for human interaction flow, where the human interaction flow in a discussion session is represented as a tree. Proposed system extends an interactive tree based pattern mining algorithm in two ways. First, it is proposed a mining method to extract frequent patterns of human interaction to support several categories of meeting. Second, it is explored modified embedded tree mining for hidden interaction pattern discovery. Modified Embedded sub tree mining is the generalization of induced sub trees, which not allow direct parent child branches, also considers ancestor-descendant branches. The experimental results show the discovered patterns can be utilized to evaluate a meeting discussion (debate) is efficient and compare the results of different algorithms of interaction flow.