| Title |
The analysis of dynamics of women’s and men’s wages |
| Translation of Title |
Moterų ir vyrų darbo užmokesčio dinamikos analizė. |
| Authors |
Vilkončiūtė, Amanda |
| Full Text |
|
| Pages |
48 |
| Keywords [eng] |
Daugiamatė regresija, OachakosBlinderio dekompozicija, Kvantilinė regre sija, Atsitiktinių miškų analizė, XGBoost, Atramos vektoriaus regresija (SVR), Lyčių darbo užmokesčio skirtumas/ MultivariateRegression,Oaxaca–BlinderDecomposition,QuantileRegression,Ran domForest, XGBoost, Support Vector Regression (SVR), Gender pay gap. |
| Abstract [eng] |
The gender pay gap remains a persistent issue in the modern labor markets, including Lithua nia. The main purpose of this study is to analyse the gender pay gap in Lithuania, based on the most recent statistical wage survey data and to apply regression and predictive machine learning meth ods. The analysis consists of three main parts. The first part, which was dedicated to understanding the gender pay gap, revealed that the multivariate Regression explains about 48.6 percent of wage variation, meaning that more than half of the variation remains unexplained, while Oaxaca–Blinder Decomposition shows that only 19.1 percent of the gender wage gap can be explained by character istics. In addition, Quantile Regression confirms that gender pay gap exists across the entire wage distribution and increases at higher wage levels. In the second part we predicted induvial hourly wages, using machine learning models. Among the evaluated methods, XGBoost achieved the best predictive performance, with a coefficient of determination of R2 = 0.97. The last part of analysis was dedicated to predicting average occupational wages, for women and men separately, based on occupational structural characteristics. Best performing model for men was Random Forest with R2 =0.93, while for women was XGBoost R2 = 0.89. By integrating causal regression with modern machine learning models, this study demonstrates the complementary value of these approaches in analyzing and predicting gender wage inequality. |
| Dissertation Institution |
Vilniaus universitetas. |
| Type |
Master thesis |
| Language |
English |
| Publication date |
2026 |
