Study on seismic magnitude prediction based on combination algorithm

• Published in: 2017 9th International Conference on Modelling, Identification and Control (ICMIC) pp. 539 - 544
• Main Authors: Zhou, Wan-zhen, Kan, Jing-sen, Sun, Shuo
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.07.2017
• Subjects: Biological neural networks; combination algorithm; Data mining; Data models; Earthquakes; magnitude prediction; neural network; Neurons; Prediction algorithms; Predictive models; support vector machine; Support vector machines
• DOI: 10.1109/ICMIC.2017.8321703

Earthquake as a kind of natural disaster, it has a great destruction, but the magnitude of earthquake is closely related to the latitude, longitude and the depth of focal. It is very important to realize the magnitude prediction by data mining. In order to reduce the damage degree of earthquake to human's life and enhance the prediction ability of magnitude at different sites and depth of source, the support vector machine (SVM) and neural network algorithm are used to construct the magnitude prediction model based on existing seismic data, and the prediction of possible earthquake magnitude at different locations and focal depth is realized based on combination algorithm of support vector machine and neural network. Experiment results shows that the predictive ability of combination algorithm is better than using traditional SVM or neural network obviously, and have large extent overcome the disadvantage of support vector machine in solving multi-classification problems and bad selection of artificial neural network parameters easily lead to over-fitting or under-fitting and other disadvantages.