Automatic assessment of creativity in heuristic problem solving based on query diversity

Abstract

 Research, development and innovation are the pillars on which 
companies rely to offer new products and services capable of 
attracting consumer demand. This is why creative problemsolving emerges as one of the most relevant skills of the 
21st century. Fortunately, there are many creativity training 
programs that have proven effective. However, many of these 
programs and methods base on a previous measurement of 
creativity and require experienced reviewers, they consume 
time for being manual, and they are far from everyday 
activities.
In this study, we propose a model to estimate the creative 
quality of users’ solutions dealing with heuristic problems, 
based on the automatic analysis of query patterns issued during
the information search to solve the problem. This model has 
been able to predict the creative quality of solutions produced
by 226 users, reaching a sensitivity of 78.43%. Likewise, the 
level of agreement among reviewers in relation to the creative 
characteristics is evaluated through two rubrics, and thereby, 
observing the difficulties of the manual evaluation: subjectivity 
and effort.
The proposed model could be used to foster prompt detection 
of non-creative solutions and it could be implemented 
in diverse industrial processes that can range from the 
recruitment of talent to the evaluation of performance in 
R&D&I processes.