Assessing The Level of Manufacturing Value Added of G-20 and Its Relation to Innovation Inputs and Outputs

Abstract

The primary objective of this study is to forecast the manufacturing value added levels of G-20 countries by leveraging the fundamental dimensions extracted from the Global Innovation Index and ascertain the extent to which innovation indicators contribute to variations in manufacturing value added. The Random Forest algorithm, known for its versatility and precision in dealing with complex datasets, has been employed as a prominent machine learning technique to predict the manufacturing value added levels of G-20 countries during the period 2013-2022. The MVA levels of G-20 countries, obtained using average and standard deviation, were predicted with a 54.14% error rate through the assistance of innovation input and output indicators. The level predicted with the highest accuracy is the one closely aligned with the average. This study's uniqueness lies in its utilization of the Random Forest algorithm to predict value added levels based on innovation inputs and outputs, which constitute the fundamental dimensions of the Global Innovation Index.