EARTHQUAKE PREDICTION ANALYSIS USING NEURAL NETWORK
ABSTARCT :
The recent scientific advances in understanding the hierarchical nature of the lithosphere and its dynamics based on systematic monitoring and evidence of its space-energy similarity at global, regional, and local scales did result the design of reproducible inter mediate term middle-range earthquake prediction technique. The real-time experimental testing aimed at prediction of the largest earthquakes worldwide from 1992 to the present proved statistically a possibility of practical earthquake forecasting although of limited precision. In the first approximation, an accuracy of 1-5 years and 5-10 times the anticipated source dimension is achieved. Further analysis of seismic dynamics allows reducing the spatial uncertainty down to 1-3 source dimensions, although at the cost of additional failures-to-predict. Despite of limited accuracy a considerable damage could be prevented by timely knowledgeable use of the existing predictions and earthquake prediction strategies. The link of theoretical research in modeling earthquake sequences in frames of statistical physics on the one hand and instrumental and algorithm developments on the other hand help developing a new generation of earthquake prediction technique of higher accuracy.
EXISTING SYSTEM :
The architecture of the neural network model for predicting the earthquake occurrence. The input neuron layer has seven nodes representing the seven seismicity indicators. They are used to recognize the pattern of earthquake occurrences. It can also be employed for predicting the nature of the future events and as well as mitigation of earthquake risk. The learning method in this paper is back propagation. It consists of two main phases: propagation and weight update
DISADVANTAGE :
The problem of earthquake prediction as a classification task where the magnitude ranges of the largest seismic event in a pre-defined time window (for instance, 1 month) are the output classes. So, the proposed methods are used to predict the magnitude of the biggest earthquake (within 0.5) in a pre-defined region in the following month
The logical consequence was paper published in Neural Networks in 2009. In this paper, the authors proposed the architecture of a probabilistic neural network (PNN) as a solution for the same problem that was formulated in .Adeli and Panakkat also used the same set of seismicity indicators as input data for training the network.
The problem of earthquake prediction was treated as a regression task: four regressors (generalized linear models, gradient boosting machines, deep learning and random forest) and ensembles for them were applied.
PROPOSED SYSTEM :
We present performance evaluation for different configurations and neural network structures that show prediction accuracy compared to other methods. The proposed scheme is built based on feed forward neural network model with multi-hidden layers. The model consists of four phases: data acquisition, pre-processing, feature extraction and neural network training and testing. In this study the neural network model provides higher forecast accuracy than other proposed methods. Neural network model is at least 32% better than other methods. This is due to that neural network is capable to capture non-linear relationship than statistical methods and other proposed methods.
ADVANTAGE :
We propose a new neural network model to predict earthquakes in northern Red Sea area. Although there are similar models that have been published before in different areas, to our best knowledge this is the first neural network model to predict earthquake in northern Red Sea area. We analyze the historical earthquakes data in northern Red Sea area for different statistics parameters such as correlation, mean, standard deviation, and other. We present different heuristic prediction methods and we compare their results with our neural network model. Details performance analysis of the proposed forecasting methods shows that the neural network model provides higher forecasting accuracy.
|