Prediction of Compressive Strength in High Performance Concrete with Hooked-End Steel Fiber using K-Nearest Neighbor Algorithm
Keywords:High Performance Concrete, Steel Fiber Reinforced Concrete (SFRC), Compressive Strength, Strength Prediction, K-Nearest Neighbor Algorithm (KNN), Machine Learning, Lazy IBK.
AbstractIn this study, the predictive capability for compressive strength of IBK a K-Nearest Neighbor algorithm was put to test in High Performance Concrete (HPC) with steel fiber addition. To achieve this objective, 150 x 300 mm cylindrical specimens were casted at least three for each batch and steel fibers were added from 0.50% - 2.00% at 0.25% interval. The mean and standard deviation were determined, and these were used to generate 100 compressive strength values within this range for each proportion. IBK classifier with K =1 nearest neighbors and 3 split percentages for training and testing were utilized. Results indicate that it is possible to generate good compressive strength results from good mean and standard deviation values. The prediction capability was very high using this algorithm with small amount of associated errors. Validation of the model using predicted versus actual results shows a very high correlation coefficient. This result indicates the efficiency of the model and its predictive capacity. It also indicates that this can improve the optimization capacity of HPC mixtures with steel fiber addition.
How to Cite
Open access licenses
Open Access is by licensing the content with a Creative Commons (CC) license.
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.