Next Generation of Data Mining最新文献

{"title":"Analysis of Social Networks and Group Dynamics from Electronic Communication","authors":"Nishith Pathak, Sandeep Mane, J. Srivastava, N. Contractor, M. S. Poole, Dmitri Williams","doi":"10.1201/9781420085877.ch19","DOIUrl":"https://doi.org/10.1201/9781420085877.ch19","url":null,"abstract":"The field of social network analysis evolved from the need to understand social relationship and interactions within a group of individuals. Knowing all individuals (employees) in an organization is difficult for an employee due to his/her limited bandwidth. Thus, in an organization’s social network, not everyone directly knows (or interacts) with each other (Cross et al., 2002). Nor does an individual observe all the communication between individuals known (or unknown) to him/her directly. The result is that each individual forms perceptions about communication between other individuals, and uses them in his/her daily tasks. Having correct perceptions for all individuals in the organization is of utmost importance for the proper functioning of business processes. Cognitive analysis of social networks has grown out of this interest in understanding what an individual's perceptions are about other individuals in terms of who they know (socio-cognitive network analysis), or what knowledge they have (cognitive-knowledge network analysis) (Wasserman and Faust, 1994). Traditional cognitive analysis approaches depend on the use of surveys and feedback from individuals. However, the lack of inability to collect large datasets, as well as problems such as inherent bias in responses, makes it difficult to analyze such social networks on a large scale. The widespread adoption of computer networks in organizations and the use of electronic communication for business processes have fostered a new age in social network analysis. E-mail communication, for example, is widely used by employees to exchange information. An email server observes all such communication between individuals in the organization, and therefore can analyze the email logs to determine the perceived social network for each individual, as well as the gold standard (or ground truth or real) social network. Given the large dataset sizes, it is difficult to apply existing techniques, since they do not scale very well. Hence, new efficient, scalable techniques are required for the socio-cognitive network analysis. First, the problem of socio-cognitive analysis of a social network is presented. This is described using email communication network, and then our previous simple yet scalable approach is presented for such analysis. The approach can likewise be applied to other communications like instant messages. Previous case study using the proposed approaches on Enron email logs is then described. It uses the Enron email dataset, wherein the email communication between the employees of Enron is analyzed using the email logs before and after the Enron crisis of 2001. The second part of the paper describes the problem of modeling and analysis of group dynamics in a social network. Data logs from a multi-player network based game, Sony EverQuest2, are now available, and are part of our current research on group dynamics. A brief overview of this problem is described and current research directions ","PeriodicalId":406196,"journal":{"name":"Next Generation of Data Mining","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1900-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134567694","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}