Presenting a Model for Predicting CBR and UCS of Expensive Soil Stabilized with Hydrated Lime Activated with Rice Husk Ash Using the Hybrid MARS-EBS Method

California bearing ratio is one of the most important design parameters of flexible pavements and unconfined compressive strength of soil is one of the important design and engineering parameters. Determining the value of these parameters through testing is time-consuming and expensive, and therefore obtaining them through alternative and reliable solutions is required. In this study, the Multivariate adaptive regression splines (MARS) is used to predict the value of CBR and UCS of expensive soil stabilized with hydrated lime activated with rice husk ash. The database used in this research includes 121 data, 70% of which are selected as training data and 30% as test data. Four input parameters of percentage of hydrated lime activated with rice husk ash, plastic limit, plastic index and maximum dry density are used in the prediction model of CBR. Also, for the UCS prediction model, five parameters of HARHA, PL, PI, OMC and MDD have been used as input parameters. Which shows that in this study, more limited input variables were used to model these two parameters compared to the models developed by researchers in the past.The value of R2 for CBR model is equal to 0.9995 and 0.9994 and for UCS model is equal to 0.9997 and 0.999, respectively, based on training and test data, which indicates the appropriate accuracy of the developed models. Also, the results of ANOVA test showed that the percentage of HARHA has the highest degree of importance for predicting CBR and UCS.