Articles

Cross-Country Transfer Learning for FDI Forecasting in Small Macroeconomic Datasets

/in /by

This study examines whether transfer learning can improve machine learning performance under data-scarce forecasting conditions. The empirical application focuses on forecasting foreign direct investment (FDI) inflows using CatBoostRegressor. The source domain is represented by a broad international panel of countries, while the target domain consists of Central Asian economies.

CatBoost is evaluated in two settings: without transfer learning, where the model is trained only on target-region data, and with transfer learning, where the model is first pretrained on the source domain and then fine-tuned on the target domain. The results are compared with simple baseline models, including a lagged-target naive model and a regional mean benchmark.

The findings show that transfer learning substantially improves predictive accuracy. Compared with the target-only CatBoost model, the transfer learning model reduces RMSE from 4.0421 to 3.0426 and MAE from 3.3504 to 2.3369. also improves from − 5.8205 to − 0.0761. These results suggest that transfer learning can help stabilize machine learning forecasts when target-domain observations are limited.

Explainable AI for Foreign Direct Investment Analysis: Evidence from Central Asia

/in /by

Foreign direct investment (FDI) is an important factor in the economic development of Central Asian countries, where investment flows have traditionally been concentrated in resource-based sectors. In the context of a growing focus on diversification, the need to analyze and study the determinants of FDI is increasing.

This study examines the determinants of FDI inflows in Central Asian countries using machine learning methods (CatBoost) and explainable artificial intelligence (SHAP), and compares the results with a classical econometric approach based on a two-way fixed effects (TWFE) model. Given the limited availability of data, a transfer learning approach is applied: the model is first trained on a group of countries structurally similar to Central Asia and then fine-tuned on the regional sample.

The results show that key macroeconomic factors such as Trade (% of GDP), Current account balance (% of GDP), and several other macroeconomic variables remain significant across both methodologies. At the same time, ML identifies additional regional patterns, such as a higher importance for FDI of determinants including Adjusted savings: carbon dioxide damage (% of GNI), Urban population (% of total population), and Access to electricity (% of population), among others.

The findings indicate that XAI provides interpretable results that are consistent with classical methods and additionally allows for capturing nonlinearities and regional heterogeneity. The study extends the application of ML and XAI in data-constrained Central Asian settings and demonstrates the value of combining econometric and machine learning approaches in the analysis of FDI determinants.