Fuzzy logic classification of seismic events

Two different methods that incorporate fuzzy logic are examined for the classification of seismic events from time-series data. Linguistic descriptors incorporate characteristics such as energy content, clutter, and bursts along with well developed heuristic rules as part of a classification scheme. A direct approach involves preprocessing the waveforms and applying the parametric data to a fuzzy associative memory. In the case of direct application, the test database was capable of 100% training of three event classes from a single seismic station.