Machine Learning-based Scaling Sensitive Feature Selection in Producing Wells
LI Chunsheng
ZHANG Dazhou
Abstract：The selection of producing well scaling characteristics plays an important role in scaling prediction.Based on ma-chine learning knowledge,this paper studies the scale formation characteristics of produced wells of an oilfield enterprise,and takes these as data samples,applies Lasso algorithm to select sensitive scale formation characteristics,and uses support vector ma-chine,random forest and decision tree for scale classification and prediction.The validity of the sensitive features is evaluated by four indicators,including accuracy,accuracy,recall and F1 value.The experimental results show that the features selected by Las-so are superior to the original feature data set in the four indexes of the classifier,which indicates that the Lasso algorithm has a good effect on the selection of scale characteristics in producing wells.
Keywords：producing well scalingfeature selectionLasso algorithm
Publication Date：2025-11-20
Online Publishing Date：2026-01-28(First online date of this platform, not the publication date of the document)
Pages：6( 3104-3108,3246 )