To improve the accuracy of the extrapolated VS30 values of boreholes on engineering sites with shallow overburden

this paper selects certain borehole data from KiK-net database for research. The correlation coefficient matrix was adopted to test the collinearity of parameters

and three parameters that can reflect the characteristics of borehole profiles were determined. Through the optimized subset method

the relationship functions between the residuals of VS30 estimated by bottom constant velocity (BCV) model and the three selected parameters were established. On the basis of three feature selecting methods including adjusted R2

Bayesian information criterion

and k-fold cross-validation

the optimal function was proposed. According to the distribution of mean and standard deviation of residuals between predicted and observed VS30 values

a correction function of BCV model was proposed. Results show that the extrapolation model of VS30 could improve the prediction accuracy of BCV model

and it was applicable in Xinjiang region. The method proposed in this paper can provide reference for developing empirical models for estimating VS30 in other regions of China.