Handwritten Digital Image Classification Based on PCA Dimensionality Reduction

Dimensionality reduction is an important idea in machine learning. In the machine learning, some high-dimensional data sets are often encountered. In the case of high-dimensional data, data samples are sparse and distance calculations are difficult. Such problems are serious problems faced by all machine learning methods. "Dimensional disaster." In addition, linear correlation between features is easy to occur in high-dimensional features, which means that some features are redundant.