Applying machine learning in predicting gross domestic savings : the case of Finland
Doan, Thi Thanh Thanh (2020)
Pro gradu -tutkielma
2020
School of Business and Management, Kauppatieteet
Kaikki oikeudet pidätetään.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi-fe2020061844978
https://urn.fi/URN:NBN:fi-fe2020061844978
Tiivistelmä
The main objective of this quantitative study is to predict gross domestic savings in Finland with the help of machine learning algorithms. Machine learning-based models and the traditional time-series model were assessed by model estimation and model performance evaluation to answer the question of whether machine learning models can yield more accurate forecasts than the simpler time-series models.
The thesis was conducted in Helsinki, Finland, during winter 2019 and spring 2020. Secondary data was gathered and collected from three different statistic sources in Finland.
The empirical analysis identified that the critical factors of the domestic saving rate in Finland during the period of Q1 1995-Q4 2019 are financial and income variables. All machine learning models produced forecasts that had a lower prediction error than the traditional linear model. In short, the empirical results indicated that by applying machine learning models, better predicting performance could be achieved in terms of gross domestic saving.
The thesis was conducted in Helsinki, Finland, during winter 2019 and spring 2020. Secondary data was gathered and collected from three different statistic sources in Finland.
The empirical analysis identified that the critical factors of the domestic saving rate in Finland during the period of Q1 1995-Q4 2019 are financial and income variables. All machine learning models produced forecasts that had a lower prediction error than the traditional linear model. In short, the empirical results indicated that by applying machine learning models, better predicting performance could be achieved in terms of gross domestic saving.