Machine Learning Techniques For Assessing Economic Factors Affecting Foreign Direct Investment Trends In Pakistan

Abstract

The present research explores the complex realm of foreign direct investment (FDI) in Pakistan by analyzing FDI trends from 1997 to 2021 using sophisticated Machine Learning (ML) techniques including K Nearest Neighbors (KNN), Support Vector Regression (SVR), and Random Forest (RF). The research analyzes the effects of economic variables like trade openness, interest rate and rate of inflation on FDI inflows and evaluates the effectiveness of these cutting-edge machine-learning techniques, employing a rich tapestry of data from the World Bank, State Bank of Pakistan, [1]and World Development Indicators. A strong explanatory power is noticeable from the regression model's which explains approximately 59.7% of the dependent variable's fluctuation. Findings of the study reveal that KNN and RF happen to be the most dependable models as compared to SVR having lowest accuracy. The study also highlights how important these economic factors are in determining FDI trends. Further, the “Trade openness” (To) feature is found to be the most influential based on feature importance values. Therefore, in order to improve FDI in Pakistan, it is crucial that stakeholders pay attention to these insights.