Abstract: In order to evaluate expressway slope stability rapidly and accurately based on a small amount of indoor and field tests during initial survey, this paper presents a new slope stability assessment method. Based on slope data from Fujian expressways, field slope stability condition was investigated, and 14 factors that might affect the slope stability were collected; then the Logistic stepwise regression was used to select significant factors, and finally a slope stability evaluation Logistic regression model with significant factors as independent variables and field slope stability condition as dependent variable was established. This model meets goodness-of-fit test, it is very stable, and its predictive results are very accurate. Moreover, the factors investigated are relatively easy to obtain; the model can handle qualitative variables; it can also assess the slope stability conveniently and rapidly. Thus, the Logistic regression method for slope stability evaluation can be widely applied to expressway slope engineering.

Key words: Logistic regression, expressway slopes, slope stability evaluation, impact factor, dependent variable