Journal of Intellectual Property (J Intellect Property; JIP)

KCI Indexed
OPEN ACCESS, PEER REVIEWED

pISSN 1975-5945
eISSN 2733-8487
Research Article

Automatic Classification of Non-Patent Literature via Patent-Literature Text Mining

1Master’s Student, Department of Intellectual Property Convergence, Gyeongsang National University, Republic of Korea
2Professor, Department of Management Information Systems, Gyeongsang National University, Republic of Korea
3Professor, Department of Computer Science, Gyeongsang National University, Republic of Korea

Correspondence to Suwon Lee, E-mail: leesuwon@gnu.ac.kr

Volume 19, Number 2, Pages 117-141, June 2024.
Journal of Intellectual Property 2024;19(2):117-141. https://doi.org/10.34122/jip.2024.19.2.6
Received on February 01, 2024, Revised on March 14, 2024, Accepted on May 29, 2024, Published on June 30, 2024.
Copyright © 2024 Korea Institute of Intellectual Property.
This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives (https://creativecommons.org/licenses/by-nc-nd/4.0/) which permits use, distribution and reproduction in any medium, provided that the article is properly cited, the use is non-commercial and no modifications or adaptations are made.

Abstract

To file a patent or examine a submitted patent, one must perform a prior-art search that includes both patent and non-patent literature. Unlike patent literature, non-patent literature is not standardized and lacks a unified search system, thus necessitating separate searches for patents and non-patents. This renders the process particularly challenging for the latter. Hence, classification methods used in patent literature are applied to non-patent literature in this study, thus enabling a search system that operates in the same manner as patent-literature searches. The proposal includes the application of machine-learning techniques to recommend or automatically assign patent-classification codes to non-patent literature. For example, a process is reviewed in which international patent classificationcodes are automatically assigned to scholarly papers using machine-learning algorithms. Based on analyzing methods that leverage text-similarity and text-classification algorithms, the automatic classification of non-patent literature through patent-literature text mining is shown to be effective and thus warrants further research. Building a database of non-patent literature coded with patent classifications can result in a more efficient prior-art search process by allowing searches under a unified classification system for both patent and non-patent literatures.

Keywords

Patent Literature, Non-Patent Literature, Text Mining, Automatic Classification, Text Similarity, Text Classification

Notes

Conflicts of Interest

No potential conflict of interest relevant to this article was reported.

Funding

The author received manuscript fees for this article from Korea Institute of Intellectual Property.

Section