This study aims to utilize principal component analysis (PCA) to analyze the effects of various chemical components on the behavior of cementitious composites containing cellulose nanomaterials and basalt fiber pellets (BFP). The PCA approach, representing nearly 80% of the data set’s variance, effectively identified several patterns among the materials. However, the dispersion along the second principal component (PC2), influenced by water (H2O), nanosilica (NS), nanofibrillated cellulose (NFC), and nanocrystalline cellulose (NCC), limits the conclusions that can be drawn. To refine the analysis, a second PCA was conducted, excluding variables with high orthogonality, such as H2O, NS, NCC, and NFC. This refinement increased the total variance explained by 10% and allowed a clearer separation of mechanical properties influenced by BFP and fine aggregates. BFP showed a high correlation with key mechanical properties, such as initial setting time (IST), final setting time (FST), residual strength, toughness (T), and Ri, due to its inherent properties. Fine aggregates were correlated with first peak strength, ductility factor (DF), and compressive strengths at 3 and 28 days (Fc′3d and Fc′28d). The refined PCA approach demonstrated that excluding certain variables highlighted the mechanical features influenced by BFP and fine aggregates, which were not evident in the initial PCA. This method, termed PCA-driven PCA, proved effective in revealing more relevant patterns in the data, underscoring the value of iterative PCA refinements for better understanding the behavior of cementitious composites.
Sawda et al. (Wed,) studied this question.