Abstract:

Innovations play a very important role when considering the performance of a particular country and its economy. Innovations are also the key to a higher quality of life, better jobs and economy and sustainable development. The innovation policy is also a crucial eminent element of both national and European Union strategy.  The main aim of this paper is to present an adaptation of the Poisson regression for symbolic interval-valued data. To estimate such regression the centres method is proposed and used in the empirical part. In the empirical part data concerning the number of patent applications to the European Patent Office for 31 countries (not only the EU ones) is used. The results show that all variables used in the model were meaningful and the expenditures of the higher education sector on research and development (as the share of the GDP) have the most significant impact on the number of patents. When using the ensemble approach with the median and the mean, the proposed approach reaches quite good results in terms of the Brier score.