SUMMARY

The possibility of using neural network for intermediate earthquake prediction

(Example in West Bắc Bộ, Việt Nam)

Cao Đình Trọng, Cao Đình Triều

In this paper, the authors set up the algorithm of analyzing input data, choosing the neural network and time duration for the intermediate term earthquake prediction (example in West Bắc Bộ, Việt Nam). The results have shown that:

- For effectively using neural network in the intermediate term earthquake prediction, the following remarks can be paid attention: the analyze of input data (from earthquake catalogue); selection of the neural network and time duration in the prediction for calculating the specific minimum average magnitude and Gutenberg-Richter function.

- Examples about earthquake prediction in the West Bắc Bộ region show that the earthquake catalogue from 1976 to 2011 is the better data for intermediate earthquake prediction by the neural network; the minimum average magnitude is 3.5; FBP neural network is most suitable for intermediate term and the time duration of prediction for magnitude 5.0-6.0 is 3 years and for magnitude 6.0-7.0 is 4 years.

Người biên tập: PGS.TS Lê Hải An.