Europos šalių investicinio patrauklumo vertinimas dirbtiniais neuroniniais tinklais
This project focuses on the topic of Europe countries foreign direct investment modelling by artificial neural networks. Foreign direct investment is significant for each country, so forecasting can help to plan the country's budget. In the scientific literature, only the factors that determine...
Main Author: | |
---|---|
Other Authors: | , |
Format: | Master Thesis |
Language: | Lithuanian English |
Published: |
Institutional Repository of Kaunas University of Technology
2020
|
Subjects: | |
Online Access: | https://vb.ktu.edu/KTU:ELABAETD37898345&prefLang=en_US |
id |
ftkaunastuniv:oai:ktu.edu:elaba:37898345 |
---|---|
record_format |
openpolar |
spelling |
ftkaunastuniv:oai:ktu.edu:elaba:37898345 2023-07-16T03:59:14+02:00 Europos šalių investicinio patrauklumo vertinimas dirbtiniais neuroniniais tinklais Evaluation of investment attractiveness of European countries by artificial neural networks Lukauskas, Mantas Ruzgas, Tomas Bruneckienė, Jurgita info:eu-repo/date/embargoEnd/2020-06-01 application/pdf https://vb.ktu.edu/KTU:ELABAETD37898345&prefLang=en_US lit eng lit eng Institutional Repository of Kaunas University of Technology https://epubl.ktu.edu/object/elaba:37898345/37898345.pdf https://vb.ktu.edu/KTU:ELABAETD37898345&prefLang=en_US info:eu-repo/semantics/embargoedAccess artificial neural networks foreign direct investment recurrent neural networks long-short term memory neural networks machine learning info:eu-repo/semantics/masterThesis 2020 ftkaunastuniv 2023-06-26T22:10:14Z This project focuses on the topic of Europe countries foreign direct investment modelling by artificial neural networks. Foreign direct investment is significant for each country, so forecasting can help to plan the country's budget. In the scientific literature, only the factors that determine foreign direct investment are considered, but FDI forecasting is not performed. The analysis of the scientific literature made it possible to distinguish five main elements of attractiveness of FDI. In this analysis, a theoretical model of the groups of factors determining foreign direct investment was formed, consisting of 16 groups of factors (total of 41 factors). The study uses models of artificial neural networks: RNN, LSTM, GRU, and ELM, which are discussed in more detail in the description of methods. In this study, more than 140,000 models of artificial neural networks were created. The results of these models have shown that maximum accuracy is achieved using ELM techniques. Comparison of the developed models with the linear regression models used by other scientists confirmed that the models of artificial neural networks are more accurate in forecasting foreign direct investment. The best models of artificial neural networks applied showed that, compared to 2017, in 13 out of 29 countries used in the study in 2018, foreign direct investment is projected to decline. In the remaining 16 countries, based on best forecasting models, FDI growth is projected. The largest increase in FDI (considering the percentage change in FDI) is expected in Iceland - 282%. Meanwhile, the largest decline in FDI (considering the percentage change in FDI) is expected in Finland. It is also worth noting that the largest FDI country in the Netherlands, the increase in FDI will be only 0.18%. Arranged models for the classification of artificial neural networks have confirmed that inflation rates in the country, export growth rates, return on equity, and so on are of the greatest importance to positive FDI. The study identified twenty ... Master Thesis Iceland KTU ePubl (Kaunas University of Technology) |
institution |
Open Polar |
collection |
KTU ePubl (Kaunas University of Technology) |
op_collection_id |
ftkaunastuniv |
language |
Lithuanian English |
topic |
artificial neural networks foreign direct investment recurrent neural networks long-short term memory neural networks machine learning |
spellingShingle |
artificial neural networks foreign direct investment recurrent neural networks long-short term memory neural networks machine learning Lukauskas, Mantas Europos šalių investicinio patrauklumo vertinimas dirbtiniais neuroniniais tinklais |
topic_facet |
artificial neural networks foreign direct investment recurrent neural networks long-short term memory neural networks machine learning |
description |
This project focuses on the topic of Europe countries foreign direct investment modelling by artificial neural networks. Foreign direct investment is significant for each country, so forecasting can help to plan the country's budget. In the scientific literature, only the factors that determine foreign direct investment are considered, but FDI forecasting is not performed. The analysis of the scientific literature made it possible to distinguish five main elements of attractiveness of FDI. In this analysis, a theoretical model of the groups of factors determining foreign direct investment was formed, consisting of 16 groups of factors (total of 41 factors). The study uses models of artificial neural networks: RNN, LSTM, GRU, and ELM, which are discussed in more detail in the description of methods. In this study, more than 140,000 models of artificial neural networks were created. The results of these models have shown that maximum accuracy is achieved using ELM techniques. Comparison of the developed models with the linear regression models used by other scientists confirmed that the models of artificial neural networks are more accurate in forecasting foreign direct investment. The best models of artificial neural networks applied showed that, compared to 2017, in 13 out of 29 countries used in the study in 2018, foreign direct investment is projected to decline. In the remaining 16 countries, based on best forecasting models, FDI growth is projected. The largest increase in FDI (considering the percentage change in FDI) is expected in Iceland - 282%. Meanwhile, the largest decline in FDI (considering the percentage change in FDI) is expected in Finland. It is also worth noting that the largest FDI country in the Netherlands, the increase in FDI will be only 0.18%. Arranged models for the classification of artificial neural networks have confirmed that inflation rates in the country, export growth rates, return on equity, and so on are of the greatest importance to positive FDI. The study identified twenty ... |
author2 |
Ruzgas, Tomas Bruneckienė, Jurgita |
format |
Master Thesis |
author |
Lukauskas, Mantas |
author_facet |
Lukauskas, Mantas |
author_sort |
Lukauskas, Mantas |
title |
Europos šalių investicinio patrauklumo vertinimas dirbtiniais neuroniniais tinklais |
title_short |
Europos šalių investicinio patrauklumo vertinimas dirbtiniais neuroniniais tinklais |
title_full |
Europos šalių investicinio patrauklumo vertinimas dirbtiniais neuroniniais tinklais |
title_fullStr |
Europos šalių investicinio patrauklumo vertinimas dirbtiniais neuroniniais tinklais |
title_full_unstemmed |
Europos šalių investicinio patrauklumo vertinimas dirbtiniais neuroniniais tinklais |
title_sort |
europos šalių investicinio patrauklumo vertinimas dirbtiniais neuroniniais tinklais |
publisher |
Institutional Repository of Kaunas University of Technology |
publishDate |
2020 |
url |
https://vb.ktu.edu/KTU:ELABAETD37898345&prefLang=en_US |
genre |
Iceland |
genre_facet |
Iceland |
op_relation |
https://epubl.ktu.edu/object/elaba:37898345/37898345.pdf https://vb.ktu.edu/KTU:ELABAETD37898345&prefLang=en_US |
op_rights |
info:eu-repo/semantics/embargoedAccess |
_version_ |
1771546817041793024 |