A model to recognize handwritten digits. Multiclass Logistic Regression Model (One-vs-all Classification) is been used to solve this problem. (In progress).
Problem Definition: Given a set of data of handwritten digits we develop a model to learn and also classify visual inputs of digits to their corresponding classes.
Analysis and Design: Data Collection/Preprocessing: We will be using MNIST data set for training. The original black and white (bilevel) images from NIST were size normalized to fit in a 20x20 pixel box while preserving their aspect ratio. The resulting images contain grey levels as a result. The images are centered to the center of mass of the pixels in a 28x28 image by computing.
Study and understanding of Algorithm Multiclass Logistic Regression Model (One-vs-all Classification) will be used to solve this problem. It is a model that is used to predict the probabilities of the different possible outcomes of a categorically distributed dependent variable, given a set of independent variables (which may be real-valued, binary-valued, categorical-valued, etc.).