Multifractal characterization for classification of self-affine signals

In this paper, a novel multifractal approach to the classification of unknown self-affine signals is presented as an improvement over traditional traffic signal classifiers. The fundamental advantages of using multifractal measures include normalization and a very high compression ratio of a signature of the traffic, thereby leading to faster implementations, and the ability to add new traffic classes without redesigning the traffic classifier. The variance fractal dimension trajectory is used to provide a multifractal "signature" for each type of traffic over its duration, and the modelling of its statistical histograms provides further compression and generalization. Principal component analysis is used to reduce the dimensionality of the data, and the k-means clustering algorithm is used to determine the number of classes in the multifractal signatures. A probabilistic neural network is then trained with these signatures, and its performance on classifying unknown traffic is used to indicate the most likely number of classes in the data.