Neuralnetwork

This project is an implementation of a neural network that utilizes C+++ parellelism libraries to analyze a large data set of color values. Using a confusion matrix, test and training data, perceptrons, hidden layer, OpenMP, learning rate, momentum, and other coefficients, this program guesses the numbers (1 through 10) represented within the test data and compares that against the actual values the test data is supposed to represent. Test data is the MNIST data set. The confusion matrix represents the guesses the program makes against the actual values represented.