PREDICTING THE CONCRETE COMPRESSION STRENGTH BY THE APPLICATION OF SUPPORT VECTOR MACHINES
Palavras-chave:
Concrete Compression Strength, Intelligent Systems, Support Vector Machine, Mechanics of the materials, Artificial intelligenceResumo
The compressive strength of concrete is one of the most used parameters in structural
engineering. In this study, the concrete compression strength is characterized by the presence and
quantification of the following components: cement, blast furnace slag, fly ash, water, superplasticizer
content, coarse aggregate, fine aggregate and age. This resistance is usually defined through destructive
experimental tests, a process that requires time and money. Another way to obtain this parameter is to rely
on computational intelligence techniques, especially with the use of an intelligent system. This study
proposes the determination of the compressive strength of concrete with a computer science concept entitled
support vector machines (SVM). SVMs technique is a machine learning method that has been used for
pattern recognition and regression analysis, which is the current case. The database used in the study is
obtained from the established and widely used studies by Yeh [1]. As performance parameters of the
method, MSE and R2
are used. The main objective of this investigation is to obtain an analytical equation
relating the constituents to the compressive strength of concrete by means of an intelligent system.
