Networks, obstacles, and resources for innovative performance: An analysis via neural networks for prediction in the manufacturing industry

Abstract

This document aims to predict the level of innovation in manufacturing companies in Colombia between the years 2017-2018. A forecasting mechanism for innovation performance has been constructed using Neural Networks (NNs). This model considers the objectives of innovation, the obstacles to innovation, the knowledge networks, and the technical information of each one of the firms. Results show that demand push, vertical sources, financial obstacles, and, qualified personnel are the most important variables in predicting innovative performance. Our empirical analysis uses firm-level innovation survey data from the EDIT (Encuesta de Desarrollo e Innovación Tecnológica in Spanish, Technological Development, and Innovation Survey in English) for Colombia for the years 2017-2018.