eduzhai > Applied Sciences > Engineering >

A specifically designed machine learning algorithm for GNSS position time series prediction and its applications in outlier and anomaly detection and earthquake prediction

  • Save

... pages left unread,continue reading

Document pages: 30 pages

Abstract: We present a simple yet efficient supervised machine learning algorithm thatis designed for the GNSS position time series prediction. This algorithm hasfour steps. First, the mean value of the time series is subtracted from it.Second, the trends in the time series are removed. Third, wavelets are used toseparate the high and low frequencies. And fourth, a number of frequencies arederived and used for finding the weights between the hidden and the outputlayers, using the product of the identity and sine and cosine functions. Therole of the observation precision is taken into account in this algorithm. Alarge-scale study of three thousand position times series of GNSS stationsacross the globe is presented. Seventeen different machine learning algorithmsare examined. The accuracy levels of these algorithms are checked against therigorous statistical method of Theta. It is shown that the most accuratemachine learning algorithm is the method we present, in addition to beingfaster. Two applications of the algorithm are presented. In the firstapplication, it is shown that the outliers and anomalies in a time series canbe detected and removed by the proposed algorithm. In a large scale study, tenother methods of time series outlier detection are compared with the proposedalgorithm. The study reveals that the proposed algorithm is approximately 3.22percent more accurate in detecting outliers. In the second application, thesuitability of the algorithm for earthquake prediction is investigated. A casestudy is presented for the Tohoku 2011 earthquake. It is shown that thisearthquake could have been predicted approximately 2 hours before itshappening, solely based on each of the 845 GEONET station time series.Comparison with four different studies show the improvement in prediction ofthe time of the earthquake.

Please select stars to rate!


0 comments Sign in to leave a comment.

    Data loading, please wait...