Construction vibration risk assessment of engineering projects based on nonlinear feature algorithm

Abstract

Abstract In order to study the vibration engineering project construction risk assessment, algorithms based on nonlinear characteristics, a nonlinear feature extraction local linear embedding (LLE) combined with adaptive neuro-fuzzy inference system (ANFIS) assessment of risk assessment methods have been proposed. This method is first utilized by the LLE manifold method to extract a number of sample construction vibration risk assessment factors of high-dimensional data vector of the nonlinear components. The nonlinear component is then used as the input for ANFIS evaluation method to evaluate and classify construction vibration risk samples. An example shows that this method can effectively improve the accuracy of risk assessment and reduce the error rate to less than 10%. Applying it to practical projects can provide effective decision-making information for construction managers and improve the credibility of decision-making. The identification results prove that the method in this article improves the accuracy of feature extraction and vibration risk assessment, and can as a vibration risk assessment method, be applied to the actual engineering vibration risk assessment.