DeepFeature is a methodology for feature selection in tabular data that converts 1D samples into 2D images, utilizing CNNs for advanced pattern recognition. The process often involves techniques like t-SNE for mapping and DeepInsight for transformation, applied to cancer classification in genomic studies. Learn more about this methodology in Oxford Academic .
feature selection in nonimage data using convolutional neural network PDF_06_032_2.pdf