基于LLE非线性特征的ANFIS施工振动风险评估

Construction Vibration Risk Assessment Based on LLE Nonlinear Characteristics and ANFIS

施工振动风险评估是以大数据为基础的,风险评估的结果受很多因素的影响。为了从诸多影响因素中提取关键影响因素并准确评估施工振动风险。提出了一种LLE非线性特征提取与ANFIS评估相结合的风险评估方法。该方法首先利用LLE流形方法提取众多施工振动风险评估样本影响因素组成的高维数据向量的非线性成分,然后将该非线性成分作为ANFIS评估方法的输入对施工振动风险样本进行评估分类。经实例验证该方法可以有效提高风险评估的准确性,降低错分率。将其应用到实际工程中能为施工管理者提供有效决策信息,提高决策的可信度。

The risk assessment of construction vibration is based on big data,and the result of risk assessment is affected by many factors.In order to extract key influencing factors from many influencing factors and accurately assess the risk of construction vibration.A risk assessment method combining LLE nonlinear feature extraction with ANFIS evaluation is proposed.First the LLE manifold method is used to extract the nonlinear components of the high dimensional data vectors composed of the factors affecting the vibration risk assessment samples.Then the nonlinear component is used as the input of the ANFIS evaluation method to classify the construction vibration risk samples.It is shown that through the example that the accuracy of risk assessment can be effectively improved and the misclassification rate can also be reduced.