Detekce a předvyplnění analytických os

Abstract

This thesis deals with the issue of company data in the area of accounting documents in the ERP system K2. The aim of the thesis was to design and implement a solution to enable data extraction including preprocessing and its subsequent use to detect and pre-fill analytical axes using predictive models. The work consists of two parts. The first part is devoted to the introduction of the problem, data analysis, data extraction and preprocessing options, predictions with a focus on recommender systems and methods for their evaluation. The second part is focused on the presentation of the proposed solution options including the implemented architecture, description of the proposed prediction models and their user testing along with the presentation of the results.