Description.
Researchers from the IT department of the University Pablo de Olavide, use techniques based on data mining to find out the patterns in time series of seismic origin. The purpose is to find patterns that precede an earthquake so that it is possible to foresee it in advance and thus, to be able to set a preventive or emergency plan.
In this line, the head researcher of R+D, Mr. Francisco Martínez Álvarez, has developed a mathematical system through which it is possible to foresee earthquakes within the following seven days after the first patterns appear, covering a distance of 50 km around and with an accuracy of 80%. In collaboration with the University of Seville, this system has been successfully tested on time series of existing earthquakes in the Iberian Peninsula and in Chile, a country with one of the highest rates of seismic activity in the world. They are currently working on data from Japan, another area stricken by important earthquakes.
The solution to a problem.
• Even though currently there are on the market several devices that detect an earthquake in advance, none of them is as precise as this one, neither have they the possibility to widen their development or hand in analytical data. In the University Pablo de Olavide, researchers use data mining to obtain data of seismic time series – a sequence of values observed throughout time and, consequently, chronologically ordered. The ultimate objective is to find patterns in the data registered just before a telluric movement so that it is possible to get ahead the earthquake with enough time to set a preventive or an emergency plan.
• The results of this project could be applied in a software prototype used as a system to detect earthquakes.
• The scientists analyze and foresee the earthquakes under certain given circumstances. They are able to do this by using to classic technics: Quantitative Analysis Research (QAR) and Regression Analysis. Elements such as the time between an earthquake and the next, the magnitude of the previous earthquake and the soil behavior, should all be taken into account.
Innovative aspects and competitive advantages.
• This project could be extrapolated to any kind of natural disaster as long as the needed data are provided. Therefore, it could be of high interest in the foreseeing of tsunamis after an earthquake, since the current detection system consists of subsea detectors placed in buoys that are only able to warn four hours before the tsunami arrives.
• Researchers are also considering the possibility of applying data mining to foresee typhoons. However, data should be analyzed, the methodology should also be analyzed to prove that it is possible to use it and, if so, do the appropriate modifications.
Companies that might be interested.
• Companies in the field of ICT that are interested in a software prototype to be used as a system to foresee earthquakes
• Organizations involved in the environmental area devoted to the foreseeing of natural disasters.
• Civil Protection Services.
• Services that monitor seismic data.