Automation of the coding of the economic activity of enterprises using Machine Learning methods applied to data already available within the FSO (data from surveys, descriptions in the commercial register, keywords, explanatory notes for classifications etc.) to support coding.
The General Classification of Economic Activities (NOGA) is an essential element for the FSO’s statistical production. The quality of the NOGA coding of the enterprises registered in the Business and Enterprise Register (BER) has a direct impact on the results of the structural, economic and synthetic statistics that concern enterprises. These depend on stable, monitored and quality NOGA coding of the BER units. With a view to reducing the burden on enterprises and continuously improving the coding of enterprises, the project aims to automate the allocation of economic activity codes to enterprises. In an initial phase, this will be based on information that is already available within the FSO.
Staff who are responsible for NOGA coding allocate or controll an enterprise’s economic activity code using the information that is available (survey input, economic activity description from the commercial register, etc.). This is inevitably associated with a human and subjective interpretation of information available which makes standardised coding difficult.
Creation of a tool that can automatically link the NOGA codes and the enterprises registered in the BER with a coding quality that is equal or superior to the manual coding currently carried out by the NOGA team.
Standardisation of coding and minimisation of the "interpretation" factor in the NOGA code allocation process.
Improvement of the quality of NOGA coding and consequently of the entire business statistics.