D. Karthika et al., International Journal of Advances in Computer Science and Technology, 2(8), August 2013, 139-146 139 A GRAPH-BASED INTERACTION PATTERN DISCOVERY FOR HUMAN MEETINGS D. Karthika 1 , R. RangaRaj 2 1 Research Scholar,Department of Computer Science, Hindusthan College of Arts and Science, Coimbatore, India. d.karthi666@gmail.com 2 Head, Department of Computer Science, Hindusthan College of Arts and Science, Coimbatore,India. rraj75@rediffmail.com ABSTRACT Mining Human Interaction flow in meetings or general representation of any interaction face to face to meetings is useful to identify the person reaction in dissimilar situation. Activities represent the natural history of the individual and mining methods help to analyze how person delivers their opinion in different ways. Meeting interactions are categorized as propose, comment, acknowledgement, request-information, ask-opinion, positive -opinion and negative opinion. From this Detecting semantic knowledge is significant. Existing system data mining technique to detect and analyze frequent interaction patterns to discover various types of new knowledge on interactions. An interaction flow in user is represented as tree. Tree based pattern mining algorithms was planned to analyze tree structures and extract interaction flow patterns. This work has extend an interaction tree based mining algorithm in two ways: Human interaction flow in a session extraction of the similar events with temporal data mining techniques, it extract the temporal patterns from the captured substance of time series of dissimilar meetings in specific period of time. After that a graph based mining method is proposed to extract the frequent patterns and mining the best meeting pattern. Graph-based Substructure pattern mining which discovers frequent substructures patterns from the face to face meeting not including applicant invention. It builds a new lexicographic order among graphs or tree representation and maps each graph to a unique minimum DFS code as its canonical label with human interaction pattern representation. Based on this lexicographic order adopts the depth-first search approach to extract frequent connected subgraphs proficiently. An experimental result shows that proposed Graph-based Substructure pattern mining algorithm substantially outperforms than the previous tree based mining algorithms. Keywords: Human Interaction Flow, Tree based mining, Frequent itemset mining, Graph based mining and Clustering based representation. 1.INTRODUCTION Data mining, which is an important technique for discover original information, extensively adopt in numerous fields such as bioinformatics, marketing and security. KDD is the process of discovering original patterns from large data sets concerning the methods at the grouping of artificial intelligence, machine learning, statistics and database systems [2] . The process of data mining is to extract knowledge from a dataset in a human- understandable structure. In the societal dynamics such as person interaction is the one of the most important thing for considerate, how a human’s behavior or human actions under the assembly and determining whether the assembly was well organized or not is one of the main issues in meetings. To overcome these numerous methods have been anticipated to originate the interaction of flow in the meeting at each human. Individual person interaction is one of the most significant distinctiveness of group social dynamics in meetings. We are developing a face to face meeting such as when user enters the shopping in the websites and give their opinion for is such as propose an new idea, generous comments, expressing a positive opinion, negative opinion etc., that implies user part, thoughts, or purpose toward a topic. To more understand the human activities and interfering of the human interactions in meetings, have to determine higher level of semantic information such as which interactions flow often occurs in a discussion and relationship among interactions in meetings. It will help to describe imperative a pattern of interaction [1]. In this study, data mining technique is detected and examine frequent interaction patterns at meetings. Human interaction flow in a conversation ISSN 2320 - 2602 Volume 2, No.8, August 2013 International Journal of Advances in Computer Science and Technology Available Online at http://warse.org/pdfs/2013/ijacst03282013.pdf