Topic modelling approach to knowledge depth and breadth: Analyzing trajectories of technological knowledge

Abstract

Technology assessment and planning requires that we can reliably, but indirectly, measure knowledge embedded in the organization. Operationalizing knowledge embedded into companies is increasingly challenging but also more and more relevant in the current cross-disciplinary and complex technological environment. Existing approaches for operationalizing company knowledge are based on patent data and analyzing patent classifications. These approaches have, however, significant limitations. In this study, knowledge depth and breadth is studied using full-text patent data from seven large telecommunication companies totaling 157,718 patents. The data was analyzed with Latent Dirichlet Allocation, an unsupervised learning method. The results are quantified using a technological diversity metric, showing temporal changes in companies knowledge. The result show how the operationalization of company knowledge is independent of patent count and that companies have their specific trajectory of knowledge development. The approach offers a novel method of analyzing the knowledge trajectory of a company, compared to existing patent classification based methods.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 4 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.