Maximum correntropy self-organizing map for waveform classification

The self-organizing map (SOM) is a well-known unsupervised classification method for seismicfacies analysis. The conventional SOM networks and their mutation methods are mostly based onthe Euclidean distance measurement which estimates the global similarity between the input dataand weights of neuron. Such measurement is suitable for signals with a Gaussian distribution butmay not work well for seismic data where the dominant structural features (e.g., stratigraphiclayers, faults, channels, salt bodies) typically appear some specially distributed patterns. Wetherefore propose to modify the SOM by introducing a new distance measurement criterion basedon the maximum correntropy. The correntropy is mainly used to measure the local similaritybetween variables, especially for non-Gaussian signals, and therefore is more suitable than theEuclidean distance to adaptively characterize the seismic structural similarities for seismicclassification tasks. Validations on two synthetic datasets show that the proposed method yieldsmore accurate classification results and is more robust to noise than the traditional SOM algorithm.Comparison experiments on a field data example also demonstrate that the proposed method candelineate the distribution of seismic facies more effectively and stably than the traditional one.