Signal segmentation using self-organizing maps

Segmenting signals into homogeneous regions is performed in many applications. In geophysical well-log segmentation, the data sets consist of various physical measurements made at different depths down a borehole and the task is to segment the data into geologically meaningful units. The self-organizing map is a realization of an artificial neural network and finds wide application in unsupervised classification and pattern recognition problems. The author proposes using a self-organizing map for signal segmentation and demonstrate the technique by segmenting a bore-hole log