AbstractThis study investigated the different geotechnical characteristics of soil stabilization with the addition of lime and rice husk ash (RHA) in varying amounts (3%, 5%, 7%, and 9% lime, and 4%, 8%, 12%, and 16% RHA). A series of compaction, unconfined compressive strength (UCS), California bearing ratio (CBR), pH, and X-ray diffraction (XRD) tests were performed in the laboratory to assess the impact of the considered admixtures on the stabilized soil parameters. This study also proposes the successful integration of an artificial neural network (ANN) and a Gaussian process regression (GPR) with the experimental UCS observations. The lime concentration, RHA concentration, and curing period were considered as the input parameters, and UCS was considered as the response. The proposed framework was extended to analyze the compressive behavior of treated soil on the minuscule level. Such coupled machine learning-based experimental investigation can provide deep insight into the soil behavior which otherwise would have remained unexplored due to the prohibitive nature of performing laboratory experiments on a large scale. The findings demonstrated the effectiveness of RHA in addition to lime in stabilizing the considered expansive soil. With the addition of lime and RHA mix, the UCS and CBR values of the treated soil increased significantly.