۶ فروردین ۱۳۹۹
دسته:
کد : 390

» عنوان : MARS Applications in Geotechnical Engineering Systems

نویسنده : Wengang Zhang

انتشار : Springer Singapore, Year: 2020

شابک (ISBN ) : ۹۷۸-۹۸۱-۱۳-۷۴۲۱-۰;۹۷۸-۹۸۱-۱۳-۷۴۲۲-۷

تعداد صفحات : ۲۵۰

فرمت : PDF

Description:

This book presents the application of a comparatively simple nonparametric regression algorithm, known as the multivariate adaptive regression splines (MARS) surrogate model, which can be used to approximate the relationship between the inputs and outputs, and express that relationship mathematically. The book first describes the MARS algorithm, then highlights a number of geotechnical applications with multivariate big data sets to explore the approach’s generalization capabilities and accuracy. As such, it offers a valuable resource for all geotechnical researchers, engineers, and general readers interested in big data analysis.

Introduction

Background
Many geotechnical engineering problems rely on the use of empirical methods expressed in the form of equations or design charts, to determine the response of the system to input variables, which is generally referred to as the surrogate model or metamodel (metaheuristics). This is usually because of an inadequate understanding of the physical phenomena involved in the multivariate problem, or the system is too complex to be described mathematically. A typical example is the determination of the undrained frictional resistance of piles in clay. Based on field load test data, empirical methods have been proposed in which the adhesion is related to the undrained shear strength as well as other factors such as the pile length by an empirical coefficient.

For problems involving several design (input) variables and nonlinear responses, particularly with statistically dependent input variables, regression methods are usually adopted. Regression methods are well-known mathematical tools for investigating the relationship between dependent variable and independent variable(s) (Montgomery and Peck 1992).

Among alternative regression methods, linear regression (LR) is usually preferred in many studies because of its well-established form and available computer packages. This method is based on certain assumptions which must be satisfied for valid results.

 

نظرات

نشانی ایمیل شما منتشر نخواهد شد. بخش‌های موردنیاز علامت‌گذاری شده‌اند *

*

آمار

  • 1
  • 74
  • 651
  • 3,041
  • 16,822