Data preprocessing impact on machine learning algorithm performance

Abstract The popularity of artificial intelligence applications is on the rise, and they are producing better outcomes in numerous fields research. However, effectiveness these relies heavily quantity quality data used. While volume available has increased significantly recent years, this does not always lead to results, as information content also important. This study aims evaluate a new preprocessing technique called semi-pivoted QR (SPQR) approximation for machine learning. designed approximating sparse matrices acts feature selection algorithm. To best our knowledge, it been previously applied learning algorithms. impact SPQR performance an unsupervised clustering algorithm compare its results those obtained using principal component analysis (PCA) evaluation conducted various publicly datasets. findings suggest that can produce comparable achieved PCA without altering original dataset.