Institutional Repository of Vilnius University
Mašininio mokymosi algoritmų taikymas įmonių, vykdančių MTEP veiklą, atpažinimui /
Abstract
dc:descriptionThe aim of this bachelor's thesis is to develop an accurate machine-learning (ML) model that enables the State Data Agency (SDA) to automatically identify companies engaged in research and development (R&D) using secondary features. The feature matrix was created for the entire population of companies that operate in Lithuania (139 032 companies) based on previous SDA surveys, declarations from the Innovation Agency and the State Tax Inspectorate, as well as publicly available sources. A targeted sample of 600 observations was selected from the population with R&D labels assigned manually. A highly precise ML model trained on that sample and aimed at identifying companies engaged in R&D was proposed. The study showed that the most important features for company classification were historical R&D expenditures and affiliation with the "72 – Scientific research and development" economic activity. Among the features that had a moderate influence on the model's decisions were revenue, assets, and the age of a firm. The final developed model drastically reduces manual classification time, eliminates human error, and enhances reproducibility, making it suitable for integration into SDA's annual R&D survey process and transferable to other statistical institutions.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2025
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Čižov, Maksim,
Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- lit
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD239028467&prefLang=en_US