Investigation of ANN architecture for predicting the compressive strength of concrete containing GGBFS

Assessment of the Applicability of Selected Data Mining Techniques for the Classification of Mortars Containing Recycled Aggregate

The article contains the results of selected tests of physical and mechanical properties of mortars differentiated in terms of the binder used: cement, epoxy, epoxy modified with PET waste glycolysate and polyester. Each type of mortar was modified by partial (0-20% vol.) substitution of sand with an agglomerate made from waste polyethylene. The obtained results were used to build a database of mortar properties, which was then analyzed with the use of three different techniques of knowledge extraction from databases, i.e., cluster analysis, decision trees and discriminant analysis. The average results of the properties tested were compared, taking into account the type of mortar, indicating those with the most favorable parameters. The possibilities and correctness of mortar classification with the use of the indicated "data mining" methods were compared. The results obtained confirmed that it is possible to successfully apply these methods to the classification of construction mortars and then to propose mortars with such a composition that will guarantee that the composite will have the expected properties. Both the presented method of plastic waste management and the proposed statistical approach are in line with the assumptions of the currently important concept of sustainable development in construction.