论文部分内容阅读
地下断层深度的估算是重力解释难题之一,我们试利用支持向量分类(SVC)法进行计算。使用正演和非线性反演技术,通过相关误错使检测地下断层深度成为可能。但必要有一个深度初始猜测值,而且这猜测值通常不是由重力资料得。本文我们介绍以SVC作为利用重力数据估算断层深度的一种手段。在这项研究中,我们假设一种地下断层深度可归为一种类型,SVC作为一个分类算法。为了有效地利用此SVC算法,我们基于一个正确的特征选择算法去选择正确的深度特征。本次研究中我们建立了一套基于不同深度地下断层的合成重力剖面训练集,用以训练用于计算实际的地下断层深度的SVC代码。然后用其它合成重力剖面训练集测试我们训练的SVC代码,同时也用实际资料验证了我们的训练SVC代码。
Estimating the depth of the underground fault is one of the key problems in gravity interpretation. We try to calculate it using the Support Vector Classification (SVC) method. Using forward and nonlinear inversion techniques, it is possible to detect the depth of the underground fault by means of related errors. But a deep initial guess is necessary, and the guess is usually not derived from gravity. In this paper, we introduce SVC as a means to estimate the depth of a fault using gravity data. In this study, we assume that one type of underground fault depth can be classified as SVC as a classification algorithm. In order to effectively utilize this SVC algorithm, we select a correct depth feature based on a correct feature selection algorithm. In this study, we set up a set of synthetic gravity profile training sets based on different depth underground faults to train the SVC code used to calculate the actual depth of underground faults. We then tested our trained SVC code with other synthetic gravity profile training sets, and also validated our training SVC code with actual data.