AI grey box model for alum sludge as a soil stabilizer: an accurate predictive tool

Abstract

ABSTRACT By using a grey box AI model, a comprehensive study is presented on the behaviour prediction of alum sludge as a soil stabilizer. To creat models for predicting the California bearing rtio (CBR) of alum sludge as a soil stabilizer, the study employs statistical models, including multiple linear regression (MLR) and Partial least squares (PLS), and advanced artificial intelligence, including classificatoin and regression random forests (CRRF) and classification and regression trees (CART). Results show that CRRF and CART models accurately predict CBR values better than MLR and PLS models. For predicting the behaviour of alum sludge in soil stablization, the compaction number of hammer and sludge content were the most significant parameters. Gs and optimum moisture content of soil were the least important parameters. Study results provide valuable insights into alum sludge’s behaviour as a soil stablizer, which could reduce waste and promote sustainable practice.