基于逻辑回归算法的砂土液化判别概率模型分析与应用OA

Sand liquefaction caused by strong earthquakes is receiving increasing attention due to the frequency of extreme seismic events.The liquefaction possibility assessment is the primary task in the study of sand liquefaction.In this paper,a probability assessment model is established based on the field investigation of liquefaction cases,com-bined with the knowledge of probability statistics and logistic regression algorithm.The effectiveness of the model is verified by comparing with the existing deterministic liquefaction assessment methods.Furthermore,the parameters analysis affecting the liquefaction assessment results is also conducted.The results show that the liquefaction discrimi-nation model established in this paper has a success rate of 85.70%and 82.50%for rejudging the liquefaction and non-liquefaction cases;and a success rate of 88.00%and 72.00%for the discrimination of the validation set,demon-strating a good discrimination success rate.The fine particle content,overburden stress correction factor,the correc-tion coefficient for overburden stress,and the adjustment coefficient for seismic magnitude should be applied to cor-rect case data when applying this model to assess liquefaction potential,which can improve the accuracy of sand lique-faction assessment.At the same time,the model can provide specific discrimination formulas.In the future,when new samples are incorporated,the model can be further improved by adjusting and modifying based on various parameters.