Webová aplikace pro predikci časových řad

Abstract

This bachelor thesis deals with the use of methods for preprocessing and prediction of time series. The aim of the thesis was to design and implement a web application that enables automatic time series forecasting, throughout the whole process, from data preprocessing to the creation of the prediction model. The work consists of two parts. The first part is devoted to the meaning of time series, their properties, methods of their preprocessing and prediction, possibilities of evaluation of predictions and also their use in practice. The second part focuses on the description of existing solutions, design and implementation of a web application, along with user testing and presentation of results.