Distributed feature selection in social networks

Embodiments are directed to generating feature values for selectable features using feature recombination. In one scenario, a computer system accesses a set of selectable features where each feature includes one or more feature values. The computer system encodes the accessed set of selectable featu...

Full description

Saved in:
Bibliographic Details
Main Authors HAYS CHRISTOPHER ALAN, HARTMAN BRIAN JAY
Format Patent
LanguageEnglish
Published 25.02.2014
Subjects
CALCULATING
COMPUTING
COUNTING
ELECTRIC DIGITAL DATA PROCESSING
PHYSICS
Online AccessGet full text

Cover

More Information
Summary:Embodiments are directed to generating feature values for selectable features using feature recombination. In one scenario, a computer system accesses a set of selectable features where each feature includes one or more feature values. The computer system encodes the accessed set of selectable features, so that the feature values are available for selection by various different users. The computer system generates a first population of feature values from the encoded set of selectable features and ranks each feature value in the first population of feature values based on which feature values are selected by the users. The computer system also generates a second population of feature values by recombining the feature values proportional to the ranking, and further generates subsequent populations of feature values by recombining the values across multiple iterations of expression of preference by different users. These feature values then converge on a subset of preferred feature values.
Bibliography:Application Number: US201213585578