Vizualizace učení a výstupů neuronové sítě

Abstract

The purpose of this bachelor thesis is to create application for visualisation of data of artificial neural network in web environment for possible future integration with tool for management of this artificial neural network, test this application with data from existing artificial neural network. Because of these circumstance it was decided to use HTML and JavaScript languages and libraries for graphics and graphs in JavaScript language. Application can visualise topology of the artificial neural network itself, display graphs of progress of precision and accuracy of learning of artificial neural network, display Confusion matrix and analyse input and output data.