High-dimensional data is a major challenge for a high-quality machine learning model. Feature extraction is the most common technique offered in the literature to reduce the amount of data. Moreover, relevant feature extraction leads to a more efficient and reliable classification system through machine learning (ML). In this study, the authors introduce Singular Value Principal Component Linear Discriminant Analysis (SVPC-LDA), a hybrid dimensionality reduction technique that combines the three most popular feature extraction methods: Linear Discriminant Analysis (LDA), Singular Value Decomposition (SVD), and Principal Component Analysis (PCA). The proposed technique was evaluated using Gaussian NB (Naïve Bayes), K Neighbors, Gaussian Process, Linear SVC (Support Vector Classifier), SGD (Stochastic Gradient Descent), and Passive Aggressive Classifiers on the standard chronic kidney disease dataset. In addition, the efficiency of SVPC-LDA was compared with all features and with existing PCA, SVD, ICA, and LDA methods. In addition, the effectiveness of the hybridized SVPC-LDA technique was measured on a high scale, with achieved values of 98.75% accuracy, 98.07% sensitivity, 96.55% precision, 100% specificity, and 99.03% AUC. In addition, dimensionality was reduced by 15% and RMSE by 40.60%, which is better than the techniques found in the literature.