The Optimal Selecting Technique on Well Logging Items and its Application to Identify Complex Lithology Profile

Abstract

Finding out the parameters with strong reflecting ability to construct recognition function can improve effectively identifying precision on complicatesets of geophysics data. Based on building the evenness parameter among individuals in one set and distance parameter between two sets, we perform optimal choice on well logging items of four lithology sets named as general sedimentary rock set, volcano-clastic rock set, volcano-clastic sedimentary rock set, lava set in Hailaer Base. By using optimal selected items into model clustering technique, the precision of the lithology identification is more than 85%.