Evaluation of machine learning models for predicting exports of India

Analysing economic trends and predicting future value of economic variables are key factors for assessing a country's economic outlook. This study pursues a machine learning approach to predict country's exports using other macroeconomic variables, namely, country's GDP, world's exports and real effective exchange rate (REER) by deploying various machine learning algorithms. In this study, Linear Regression, Support Vector Machine, Decision Tree and Random Forest machine learning models are applied on India's trade data for the period 1970–2019. By comparing the prediction accuracy, it is found that decision tree regression outperformed the other three algorithms and produced better results for predicting India's exports.