An exploratory analysis of Brazilian universities in the technological innovation process

Abstract

The purpose of this paper is to use hierarchical clustering (HC) and principal component analysis (PCA) for determining the key institutions and variables in a multidimensional data set to visualize Triple Helix (TH) relationships between industry, academia and government. This is a huge task, essential to better understand technological innovation, an interactive process that creates knowledge in an integrated way, reducing the number of variables. For this task we analyzed the data from eight Brazilian universities between 2008 and 2015 considering median of twelve parameters so diverse as the number of research groups; researchers; teaching staff; innovation projects in collaboration; papers; patents; technology transfer agreement; money generated from technology transfer and financing. From HC it was possible to identify four main university clusters considering all variables. PCA also shown four groups on main component mapping, in agreement with HC. The financing, the existence of innovation environments and specific innovation legislation, and the regional context explain clustering. PCA suggests that much of the data variability can be summarized in three principal components, presenting industry, academia and government interrelationships, in agreement with HC. So PCA and HC could be considered as a new view of investigation to quantify the TH, statistically mapping this model.