Using machine learning and 10-K filings to measure innovation

Article
Osuva_Nousiainen_Ranta_Ylinen_Järvenpää_2024.pdf - Lopullinen julkaistu versio - 976.89 KB

Kuvaus

© 2024 The Authors. Accounting & Finance published by John Wiley & Sons Australia, Ltd on behalf of Accounting and Finance Association of Australia and New Zealand. This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.
The purpose of this paper is to develop and validate a text-based measure of innovation using latent Dirichlet allocation on a sample of 45,409 10-K filings from US listed companies. We expect that the text-based innovation measure is associated with innovation and can be used to measure innovation for companies without patents or significant research and development expenditures. The empirical results are consistent with these assumptions, but reveal that thorough initial testing is required to ensure robustness. This study extends the research on innovation measurement and company disclosures, and provides a new method for assessing innovation using company disclosures.

Emojulkaisu

ISBN

ISSN

1467-629X
0810-5391

Aihealue

Kausijulkaisu

Accounting & Finance

OKM-julkaisutyyppi

A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä