# Further research

The final part of this chapter will establish some recommendations and ideas for further research on this topic. These recommendations are based on some of the limitations previously established, or were noted during the development of the analysis. A list of the most important ones follows:

The most important further research topic is the expansion of the framework and the experimentation with it in other research areas. As a proof of concept, this project was based on the biofuels ecosystem; however, it's certainly interesting to think of the possibilities of its application to other fields. For instance, after the design of a database one could easily use exactly the same approach.

The second area of research is related to the relation between external factors (price of sugar and oil) and the production of technological assets. To study this, this project used the Pearson correlation, which produced very interesting results. However, a relationship of this nature should be studied using a time-series analysis, in order to understand exactly how the behaviour changes (or not). The application of a time series analysis and a wide range of external factors is a very interesting research topic. From its application, one could understand exactly what moves the research field.

Although the GDP per capita and the GDP difference between countries was widely used, academia has shown that these indicators might be just the tip of the iceberg in explaining the innovation systems of countries. One possibility is to dive into this topic is the application of a very large number of indexes (GDP, HDI, alphabetisation, number of universities etc.) and to study their influence when comparing to the capability correlation matrix. This is a direct way of understanding the impact of every index. Moreover, one could mix different indexes and see which combination better explains the capability correlation matrix.

The uniqueness of countries and of their usage of term pairs served as proof of concept to a notion of “specialization” of countries in a certain area. This could obviously be expanded to periods of time, or organizations. What made this year, organization, special? The notion of uniqueness is very valuable in the context of innovation, and further work should be developed in this area.

In the study of patents and publications, the volume of terms of every type was used. A more interesting and granular analysis would be to conduct the same study but with the usage of term-pairs. As previously demonstrated, term pairs provide a richness to the analysis that term frequency simply does not deliver. Also regarding the complementary analysis, the application of more unsupervised learning techniques in order to discover other clusters is definitely an interesting area of research.

In what regards the study of the nature of collaborations, a very rudimentary analysis was carried out by the detection of the string ‘Univ’ in the name of organizations. One can envision a deeper study into the types of collaborations in the industry, which would serve even better the triple helix quantification. Here, a distinction between university-industry, industry-government, and government-university could be made. To do this, several natural language processing techniques (such as regex) would have to be used.

Last updated