Nonlinear Dimensionality Reduction Of Isomap In The Analysis Of Seismic Attribute Parameter Data

It is important to forecast combination gas for seismic attribute parameter. Thereare much more seismic attribute parameters considering holonomy and appliancewhich result to high dimensional space. Although high dimensional space containsabound information, it brings much more difficulty in calculation. In order to get abetter interpretation and prediction, we should take use of a method of reduction ofdimensions associating properties of matter.If seismic attribute parameters themselves are linear, we can get a goodconclusion with linear dimensionality reduction. Else we should take measures ofnonlinear dimensionality reduction. Isomap is geometric Algorithm with nonlinearsectional metric transformation. Its feature is with geodesic distance instead ofcommon distance. It is a improvement of MDS and submitted by Tenebaum in 2000. Itcan reveal internality of nonlinear manifold, which contains fewer parameter and highefficiency.The new method named Isomap was applied in the text and matched with MDSby WNN. We evaluated the two method of dimensionality reduction from two aspects.One is prediction and the other is error. At the same time, we also predicted withprimal seismic attribute parameter and matched it with the result of Isomap. Inconclusion, Isomap proved a good effect in the analysis of seismic attribute parameterdata.