Please use this identifier to cite or link to this item:
http://hdl.handle.net/10071/36642| Author(s): | Mamede, Ricardo P. Andrade, Matias Pinheiro, Cristina Alves, Tiago Martins, Tomás Mendes, Beatriz Paiva-Silva, João |
| Date: | Mar-2026 |
| Title: | Mapping capabilities for Smart Specialisation: an LLM-based approach |
| Pages: | 1-28 |
| Collection title and number: | DC_WP_2026_1 |
| Reference: | Mamede, R. P., Andrade, M., Pinheiro, C., Alves, T., Martins, T., Mendes, B., Paiva-Silva, J. (2026). Mapping capabilities for Smart Specialisation: An LLM-based approach (WP No. 2026/01). DINÂMIA'CET-Iscte. 10.15847/dinamiacet-iul.wp.2026.01 |
| DOI (Digital Object Identifier): | 10.15847/dinamiacet-iul.wp.2026.01 |
| Keywords: | Smart specialisation strategies Política de inovação -- Innovation policy Capability mapping Large Language Models (LLMs) EU cohesion policy |
| Abstract: | Smart Specialisation Strategies have become a cornerstone of EU Cohesion Policy, yet prioritising investment areas remains a challenge. This paper introduces a methodology for capability mapping to support strategic priority setting, integrating data from patents, scientific publications, and R&D projects, classified using Large Language Models (LLMs). Unlike approaches that focus narrowly on technologies or industries, the method maps the intersection of technological domains and fields of application, as envisaged in the foundational literature on Smart Specialisation. This intersectional perspective aims to support a more precise and policy-relevant identification of areas with transformative potential. An empirical application to the Portuguese case illustrates the potential of the methodology to inform decision-making. The method grounds the analysis in local capabilities and contexts, supporting the exploration of place-based, distinctive pathways for structural transformation. |
| Peerreviewed: | yes |
| Access type: | Open Access |
| Appears in Collections: | DINÂMIA'CET-WP - Working papers com arbitragem científica |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| workingPaper_hdl36642.pdf | 794,08 kB | Adobe PDF | View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.












