Impact of learning rate in prediction with single layer neural network
MNIST (Mixed National Institute of Standards and Technology database) is a dataset including handwritten digits and their equivalent number which are gathered by National Institute of Standards and Technology in 1998. MNIST is a good entry point to work on getting more experienced in Machine Learning, Deep Learning, and related technics as well. Images size are 28×28 and all of them are grayscale. It means that we have a matrix with size 28X28 and each cell of the matrix is the pixel value in range 0 to 255. The above image is some sample images and their given digits of MNIST dataset.
I’m interested in using neural network to predict digits from handwritten images. Here is the model of each neuron which will be used in network. [more...]