Earthquakes classification using data mining techniques

In the last ten years, the time series data mining has evolved considerably. In this period have been developed different algorithms that allow pattern matching and classification, and now they are a novel set of techniques for time series analysis. In particular we use two data mining techniques and one of classical time series analysis, for the earthquakes classification: Quiron, It is an indexing technique that allows similarity search in time series using space access methods that are implemented using an abstract data type called R-Tree. STRS, This technique allows classification, clustering and anomaly detection using symbolic representation of the time series. Recurrence Plots, This is a technique quantitative and qualitative of time series. The main idea of recurrence plots is that we can observed in a time series the realization of some dynamical process and the interaction of the relevant variables over the time. In this work we show the development and evaluation of two data mining algorithms: Quiron and STRSS for classification applied to earthquakes analysis. The earthquake that we analysis were extracted from The National Database of Strong Earthquakes of Mexico developed at National Autonomous University of Mexico.