Wisdom 2.0

Expert recognition in the virtual world

Online discussion forums have become a popular means for seeking and sharing knowledge, but they vary in their effectiveness in identifying specific expertise for problem solving.

“Unlike experts in the real world whose expertise is certified by advanced degrees and/or authoritative titles, experts in the virtual world are difficult to identify,” says associate professor of business information technology, G. Alan Wang.

G. Alan Wang
The system that Alan Wang helped develop allows users to find
the most recognized experts on a given topic.

Wang is part of a research team that developed a new expert-finding technique for online forums. The computer program evaluates expertise based on both the expert's authored documents and social status within his or her knowledge community.

Expert finding has become increasingly important in large corporations, Wang says. “Companies such as GE, Dell, IBM, KPMG, Microsoft, and Google have amassed a huge volume of data — in addition to knowledge exchanges on internal discussion forums, employee emails and other internal communications and web-based customer service interactions.”

The computer program that Wang's team developed, ExpertRank, “could be easily extended or modified to these data to help build expert databases or organizational memory systems that facilitate knowledge exchange among employees,” he says.

Electronic knowledge communities

Discussing the emergence of online forums, Wang says that professionals in a variety of industries, faced with limited expertise and resources within their organizations, have long turned to the Internet as an external source of information and knowledge.

Technologically, electronic knowledge communities have evolved from traditional listservs and newsgroups into more advanced web-based discussion forums and interactive communication systems that are rich in social media, he says.

The forums rely on members to serve as primary sources of information or knowledge. Members “share interests and voluntarily help each other to solve problems by sharing expertise across time and space,” Wang says.

“A large online community may have millions of participants and a knowledge repository of millions of text documents in the form of online postings.”

But an effective knowledge source should make accessible not only knowledge but also the sources of knowledge — experts who can either solve the problems directly or suggest other sources of information as indirect solutions — expert finding is an important tool for online forums.

“When a user has a problem and is looking for a solution, he or she may spend time browsing past relevant postings or wait for responses by initiating a new discussion,” Wang says. An expert-finding technique allows the user to choose to consult the expert members of the forum directly. “It is convenient and effective for both users who seek knowledge and those who are willing to share.”

Measuring social importance

Existing expert-finding techniques have several drawbacks, however. They have been developed mostly for organizations, where information quality is high and knowledge hierarchy is well defined.

“Information quality in online communities,” Wang says, “is considerably poorer than that in organizations.” Studies have found that the quality of information provided in an online community is often inversely related to the size of its membership.

(For example, in Wikipedia — the largest, community-based, open-access online knowledge repository — only a tiny fraction of the articles [0.09 percent, as of September 2012] met assessment criteria for information quality and qualified as featured articles.)

Moreover, most automated expert-finding techniques rely on document-based relevance to predict the expertise level of experts for a given query, Wang says, although recent studies have found that one's social importance — inferred from his or her past interactions with other members in the community social network — can also help identify experts.

Both are important expertise indicators in automated expert finding, Wang says. “The more documents an individual has authored in an expertise area, the higher the degree of expertise the individual has.”

Social importance is like an authoritative title, Wang says, indicating the experts' social status or influence in their social network. “Studies have repeatedly shown that an individual's social influence plays an important role in the perception of his or her expertise.”

Very few expert-finding systems consider both document-based relevance to a given query and the expert's social importance in the community, and Wang says ExpertRank fills this need.

The researchers tested their computer algorithm in experiments using Microsoft Office Discussion groups. It performed satisfactorily overall, Wang says, while “significantly outperforming” commonly used document-based expert finding techniques. “On average, 5 of top 10 experts recommended by ExpertRank coincide with those recognized by the community members and Microsoft personnel.”

Wang co-authored “ExpertRank: A topic-aware expert finding algorithm for online knowledge communities,” with Pamplin faculty members Alan S. Abrahams and Weiguo Fan; Jian Jiao, a doctoral student in computer science at Virginia Tech; and Zhongju Zhang, an associate professor at the University of Connecticut. Their article has been published online in Decision Support Systems.

Virginia Tech Pamplin College of Business Virginia Tech Pamplin College of Business Magazine Fall 13

Shadow for bottom of page