Pattern recognition is the art of assigning a class or category to a pattern while taking into account some of its properties (features). Let each pattern be represented by a feature vector. As with image analysis (see previous section), complex networks can then be obtained where each node correspond to a pattern, while the links are defined by similarities between patterns.
The challenging problem of identifying clusters of objects can then be related to the (also challenging) problem of finding communities in networks. At the same time, the problem of identifying the class of a given network can be formally understood as a pattern recognition problem. Sound and well-established multivariate methods can then be applied in order to answer such questions.