Aiming at the complexity and fuzziness of loess collapsible degree
an evaluated model of loess collapsibility is established with the application of the theory of fuzzy measure. The evaluation index of the model is coefficient of collapsibility
and the factors are natural specific gravity
dry bulk density
moisture
void ratio and plasticity coefficient. The point estimates of the evaluation index and measure function can be obtained by training a set of learning samples. Then the classification of the distinguished samples can be got. The practice showed that
the evaluation model functions well with high prediction accuracy and can be applied to practical engineering.