APPLICATION OF MULTIPLE LINEAR REGRESSION TO POROSITY QUANTITATIVE PREDICTION OF SANDY CONGLOMERATE BODIES IN NORTHERN DX

This paper presents the application of a method,multiple linear regression,to quantitative porosity prediction of sandy conglomerate bodies from three aspects.(1) On the basis of analyzing logging data and interpreting porosity,seismic interpretation for geological horizon is carried on to extract and optimize various seismic attributes.(2) According to the analysis on the correlations and physical meanings of seismic attributes,the total strength,total negative-amplitude and zero phase numbers are determined to be the best attribute set.(3) Through the trend analysis on seismic attribute parameters and discrete data of porosity,this study distinguishes and removes singular value impact,increases both data correlation and prediction accuracy of porosity,and fulfills quantitative porosity prediction.Results coincide with the geological law and have a high precision.Therefore,this method is effective and feasible.