Companies are increasingly interested in using social networking data to automatically determine users’ behavior patterns, but researchers say more work is needed to define what it takes for two people to be connected. “You don’t get to directly observe relationships, you get to observe communication events,” says Yahoo!’s Jake Hofman. Algorithms can infer different social network patterns based on the interpretation of communications events, making these certain networks more suitable for different circumstances. For example, a network based on relatively infrequent communications might work well for sharing tagged news items, while more frequent communications might work better for networks designed for sharing more intimate information. Incomplete information can throw off attempts to characterize social networks automatically, says Georgia Tech professor Eric Gilbert. Studying the structure of a network in greater detail can compensate for the problem of incomplete information, Gilbert says. Meanwhile, Microsoft’s Munmun De Choudhury says more research is needed to make algorithms able to better understand the nature of social connections.
View Full Article
For More Information Visit: http://www.cpccci.com
