Rule Induction Algorithm for Application to Geological and Petrophysical Data

摘要: Very large geological, geophysical, and petrophysical databases often contain multiple data types that must be interpreted for application to subsurface modeling. Significant advances in discovering complex even nonintuitive relationships could lead better predictions. There is a litany of analysis techniques used today, including cluster analysis, principal component discriminant parametric nonparametric regression, N-dimensional histograms. Regression neural networks have common their multivariate combination predictor variables. These may good at interpolating within the boundaries training data, but poor extrapolation because lack understanding underlying Alternatively, machine learning mining technologies Rough Sets hold promise finding category expressing those rule-based language. This paper presents novel rule induction algorithm derived from these machine-learning techniques, developed reservoir characterization with geological geophysical data. A set facies models systematical changing geometric features synthesized. The are coded effective permeability calculated. Rules between deducted by using proposed technique. consistence rules implemented synthesization exhibit effectivity Further second example assignment wireline logs test confirmed geologists who spend significant time trying summarize well logs. probability feature distingushibility technique supplied additional information geologist reconsider original distinction among facies.