Estimating the Lengthy Missing Log Interval using Group Method of Data Handling (GMDH) Technique

摘要

Estimation of well log properties is crucial in identifying the trends and the properties in oil and gas industry, which will enable firms to avoid problems during operation procedures. In this paper, Group Method of Data Handling (GMDH) technique is utilized to generate a generic model with superior prediction capabilities. NeuraLog software is used in order to convert the scanned image logs into digit one that can be used by MATLAB code. Group Method of Data Handling is utilized to predict the same missing interval. An aggregate of 601 field data sets were utilized to create the model. These information sets were separated into training, cross validation and testing sets in the degree of 2:1:1. Trend analyses as well as graphical and statistical tools have been utilized in order to assess the model performance.