MFOFLANN: moth flame optimized functional link artificial neural network for prediction of earthquake magnitude

Santosh Kumar Majhi, Sk Sajeed Hossain, Trilok Padhi

Abstract: It is true that the world today suffers greatly from the modernization by humans however the natural disasters in no way do less harm. In the wake of this threat, there is a need of prediction of earthquake magnitude beforehand so as to determine the severity of the disaster and create an early warning system to avoid any casualty as far as possible. To curb the menace of earthquakes, there have been significant improvements in earthquake engineering regarding the study of different seismological parameters that influence earthquakes and generate useful features from them that can be used for earthquake forecasting. In this regard, there are various organizations such as the united states geological survey, the international seismological centre that collect records about details of all the earthquakes, small shockwaves occurring across major continental regions of the world, aid various research works related to earthquake modeling. This article is aimed towards a methodology of combining a neural network model known as functional link artificial neural network with both standard machine learning algorithms and certain nature-inspired optimization algorithms so as to achieve the task of predicting earthquake magnitudes.