Early identification of cardiomyocyte dysfunction is a critical challenge for the prognosis of heart failure, particularly in the context of impaired relaxation. Myocardial relaxation relies heavily on efficient intracellular calcium (Ca²⁺) handling. During diastole, a sluggish removal of Ca²⁺ from the myocyte disrupts sarcomere relaxation, potentially leading to diastolic dysfunction. Characterizing myocardial relaxation requires analysis of both sarcomere length transients and intracellular calcium kinetics. However, developing a classifier tool that leverages a minimal set of carefully chosen features, derived from sarcomere length transients and/or calcium kinetics, to differentiate normal cells from those exhibiting impaired relaxation, remains to be explored. Cardiomyocytes were obtained from a transgenic phospho-ablated mouse model with complete ablation of phosphorylation sites Ser273, Ser282, and Ser302 on cardiac myosin binding protein-C, while cardiomyocytes from wild-type, non-transgenic mice served as a control group. We are employing principal component analysis (PCA) to mitigate the high dimensionality resulting from numerous features extracted from sarcomere length transient and calcium kinetics data. Given the presence of missing values and the desire to streamline feature selection, we quantify the importance of original features of sarcomere length and intracellular calcium kinetics by calculating contribution scores to each principal component. We hypothesize that by employing feature selection through PCA, we will achieve comparable accuracy in predictive classification task compared to utilizing the entire set of features, enhancing interpretability without sacrificing predictive performance.