Prediction of California bearing ratio of subbase layer using multiple linear regression models

This research presents the prediction of California bearing ratio (CBR) of subbase layer of roads in the Makkah area, Saudi Arabia. This prediction is based on simple tests such as sieve analysis, Los Angeles Abrasion test and the relationship between dry density and moisture content. Multiple linear regression models are investigated for the estimation of the CBR from sieve analysis, maximum dry unit weight and optimum moisture content of the subbase soils. The data were collected from Makkah roads during construction and tested at the Soil Mechanics Laboratory in the Civil Engineering Department of Umm Al-Qura University. The fitted regression models indicate strong correlations (R2=0.95) between the variables mentioned above.