Image Representation of Numerical Data-points for Classification Using Convolutional Neural Network

Higher dimensional data is a problem for classification. Researchers put forward various approaches like dimension reduction, ranking features for an optimal set, different feature extraction techniques, etc. for classification. In the current context of categorization, CNN (Convolution Neural Network) is a powerful tool, which classifies the data based on images as input. In this work, we have demonstrated a method to represent individual samples in an image, irrespective of dimension, and used CNN for classification. Further, to preserve data privacy, the images are flipped randomly (horizontal and vertical flips) for classification using CNN. The outcomes of both approaches are compared with related techniques and have shown promising results.