Page 37 -

4 Suggestions for Further Research

The data herein rely heavily upon a keyword identification of clean energy patents that
focus on late-stage innovations. Future work needs to update and make available
identification and categorization using a more general and flexible ontology comprised of
a mixture of keyword, patent classification code, and assignee-based identification that is
simultaneously inclusive of earlier-stage technologies but still accurate.

Significant work will need to be undertaken to further validate the Web-presence
methodology. It would be extremely difficult to do this for the complete patent dataset
due to the labor intensity of determining whether each patent has demonstrated
commercial value. However, future work could include other institutions’ licensed
versus unlicensed patents as a method of calibrating this method. Qualitative surveys
with active inventors will enrich this line of analysis. Lines of inquiry may include the
inventor’s qualitative assessment of the patent’s value, the method of commercialization
(start-up, license to large firm, license to non-profit, not deployed and why), and the level
of commercial success. Additional focus on refining data quality around regional,
organizational, collaborative, and institutional factors, as well as the funding sources for
the research (particularly if the research was supported by a government grant) would be
helpful. Also, further effort should be expended in the automated analysis of the content
of each type of website. A variety of schemes could be investigated [87], including
simple search term counts, dictionary look-up of appropriate terms, and semi-supervised
machine learning algorithms (at least for the patents with enough hits apply such methods
[88]).

We hope that this work will contribute to an open platform for future research by the
community of science and policy scholars. The focus in this proposal on clean energy is
just one slice of this effort. We hope to build, through the collaborative work of many
scholars, a series of integrated databases that would enable the tracing of scientific
investment from grants all the way to commercialization. Figure 11 illustrates an
idealized schematic of this goal. This effort would help accelerate the development of a
system where most grants could be traced to papers, papers to papers through citations,
papers to patents through people and citations, and patents to commercial outcomes
through people (for example, on boards of directors), licensing agreements, and the
proposed Web-based measure of commercial impact and deployment. This would be a
powerful tool for both DOE and the research community to better assess the value and
success of research and development investments. This capability will be essential as the
market quickly evolves, research investments increasingly come from the private sector,
and there is a call for increased transparency on public research expenditures. Further
development of this database and the addition of new datasets outside our control through
an open-research model will enable this tool to be broadly used to investigate innovation,
commercialization, and deployment for DOE and the public sector in general.

30