Seismic databases usually contains many parametric earthquake attributes, some of them are recorded by one observing means that is associated with one type of seismic precursors, and some are observed by a different measure to seek another type of abnormal events that are potentially related to earthquakes. In seismological study it is a very common requirement to evaluate how one type of parameter is related to another, which providing a cross-verification of seismic anomalies or an estimate of earthquake consequence with a quantitative measure. This requirement can be formulated as a knowledge discovery task which can be handled by rough analysis technology. In this study we develop a rough analysis method for investigating various relations between a set of attributes by introducing two concepts in terms of key and maxima and develop a set of rough analysis algorithms. The proposed method permits us not only to reduce non-discriminant attributes in decision tables but also to quantify relations between parametric attributes. It has been applied to analyzing the relation between earthquake magnitudes and intensities within a simplified earthquake database, demonstrating its practical value.
Bi, Y., Wu, S., Shen, X., & Guan, J. (2010). Rough Analysis for Knowledge Discovery in a Simplified Earthquake Database. In Integrated Uncertainty Management and Applications Advances in Intelligent and Soft Computing (Vol. 68, pp. 489-500). Springer. https://doi.org/10.1007/978-3-642-11960-6_45