For  Eagle and Petland eigenbehaviors are a set of characteristic vectors that represent behavioral structures and they can be used to predict human behavior with high accuracy. The authors claim that despite individual, idiosyncratic, random behavior people typically have identifiable routines. By applying a reserach methodology that collected data from 100 subjects over a period of 9 months, they were able to recover information that proved that:

(…) communities within a population’s social network tend to be clustered within the same behavior space. It seems reasonable that this type of behavioral homophily is present in a variety of social networks. It should be possible  for practitioners, using virtually any type of longitudinal behavior data, to similarly quantify the behavior space of a particular group or individual of interest using the eigenbehaviors technique described above. If strong behavioral homophily is present in the data, it should equally be possible to infer an individual’s affiliations by quantifying the individual’s distance from a community’s behavior space.

The two authors show how knowledge is socially constructured Stephen Downes comments; “groups of friends can have their own collective ‘behavior space’ which corresponds to the common behaviors of the community.”

References

Eagle, N., Pentland, A.S., 2009, ‘Eigenbehaviors: identifying structure in routine’, in “Social Networks: new perspectives” (Guest Editors: J. Krause, D. Lusseau and R. James), Behav Ecol Sociobiol (2009) 63:1057–1066, DOI 10.1007/s00265-009-0739-0

Image and paper available here

Stephen Downes commentary available here