Abstract
Innovation performance has become a central issue for being economically competitive in the new global economy. It has therefore been the subject of much systemic investigation on different levels. Special attention is given to the numerous actors, their interactions and contribution to national innovation performance. This is also the focus of the concepts pertaining to Systems of Innovation (SI). The SI have become a well-established instrument for policymakers to implement innovation policies that enhance innovation performance. The cascading evolution of SI concepts and the multiple perspectives provide a risk of bias and misleading conclusions, leading to inefficiency in using SI as a policy instrument. Further, no framework exists that permits the assessment of innovation performance and its affecting factors. Relevant literature lack a robust and structured data foundation and the scope of frameworks is confined to actor-centric or economic context-related analysis. Therefore, there are few practical guidelines for policymakers about the innovation performance of nations.Based on these weaknesses, this study aims to build an AI based model for national innovation performance investigation considering the SI boundary spanning interactions and national ambiguity within a global innovation system.
The applied research methodology is quantitative. Techniques from the machine learning field of AI are used such as Deep Neural Network, Natural Language Processing, Random Forest, Autoencoder or Local Interpretable Model Explanation on secondary data retrieved from 51 distinct databases.
The findings revealed the topics in SI overlaps and pertinent measurable indicators of actor interlinkages. A framework was developed that maps the SI and unifies existing models considering actor relations and the economic surroundings. Innovation performance related results suggest six stages referring to the ability of creating, establishing, marketing, and scaling innovation, besides innovation profiteers and leaders.
This study helps policymakers understand national innovation performance and its drivers. It shows how to create a quantifiable database from document-based data for quantitative research. The study clarifies the blurring boundaries of SI and emphasises the necessity to address national ambiguity within a generic model.
| Date of Award | 11 Dec 2023 |
|---|---|
| Original language | English |
| Awarding Institution |
|
| Supervisor | Sercan Ozcan (Supervisor), Bodo Herzog (Supervisor) & David Bacon (Supervisor) |