Porosity prediction method based on selective ensemble learning

Abstract

The invention provides a porosity prediction method based on selective ensemble learning. According to the method, a typical machine learning method is researched and analyzed; a group of individual learning models with excellent performance are selected from classic models such as a support vector machine, a radial basis function (RBF) neural network, a random forest, ridge regression and K-nearest neighbor regression through a principal component method analysis method to form an integrated learning model, wherein the weight of the individual in the integrated model is obtained by a principal component weight averaging method, and finally, the output of the integrated learning model is obtained by adopting a weighted average method. The model is called a PCA-SEN model for short. The PCA-SEN model overcomes the defects of a single model, and the generalization ability of the model is high. The reservoir porosity is predicted through the method, so that a more accurate prediction result is expected to be obtained.

Description

technical field [0001] The invention relates to the field of machine learning, in particular to a porosity prediction method based on selective integrated learning. Background technique [0002] When using machine learning methods to solve practical problems, a single learner often has shortcomings such as poor universality and low stability. In this case, ensemble learning came into being. Ensemble learning is a method to obtain better prediction performance than a single learner by combining multiple individual learners into a prediction model. At present, ensemble learning has been widely used in many fields. For classification problems, such as handwriting recognition, face recognition, image recognition, etc., ensemble learning has achieved good results; In the late period, there were relatively few research results, such as power load forecasting, financial fields, etc. But there are still some problems that need to be further solved in ensemble learning. In practic...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the problems existing in the existing porosity prediction methods, the present invention proposes a porosity prediction method based on selective ensemble learning

Images