以主題分析、隨機森林與多準則決策分析法探勘第六代行動通訊技術專利

Topic Modeling, Random Forest, and MCDM Methods Based Explorations of Patents for the Sixth Generation Wireless Communication Techniques

陳思璇(Chen Si-Syuan)

指導教授 ： 黃啟祐

國立臺灣師範大學/科技與工程學院/工業教育學系/碩士(2022年)

https://doi.org/10.6345/NTNU202201579

摘要

第六代行動通訊（6th Generation，6G）為次世代無線通訊的基礎，目前，全球仍處於探索階段，主流技術尚不明確，因此，專利發展與布局之現況，極需探索，以作為未來學術、研究與廠商產品與規格訂定之依據。雖然6G專利探勘極為重要，相關研究甚少，故本研究提出整合文字探勘探勘技術與多準則決策分析方法之新型架構，探勘6G通訊專利。
首先，本研究依據美國專利商標局（United States Patent and Trademark Office，USPTO）專利資料庫中探勘之6G通訊技術專利，透過隱含狄利克雷分佈（Latent Dirichlet Allocation，LDA）擷取主題，之後，針對每個主題，使用隨機森林迴歸（Random Forest Regression）演算法，得出每一專利主題以其他主題表示之特徵重要性後，將特徵重要矩陣轉化為決策實驗室分析法（Decision Making Trial and Evaluation Laboratory，DEMATEL）的初始影響矩陣。其次，透過基於決策實驗室之網路流程法（DEMATEL based Analytic Network Process，DANP）得出對應於每個主題的影響權重，權重最高之主題所對應之技術，為第六代行動通訊技術之驅動力量。
依據實證研究之結果，得出全息無線電技術（Holographic radio technology）、同頻同時全雙工（Co-frequency co-time full-duplex）和智能與通訊融合（Wireless artificial intelligence fusion, Wireless AI fusion）這3個關鍵技術，對6G通訊未來的發展最為重要，而全息無線電技術（Holographic radio technology）是影響力最大的技術。研究結果不僅作為瞭解6G通訊技術發展脈絡之基礎之外，也可以提供未來6G通訊技術發展參考之依據。此外，所發展之技術分析框架，也可用於探索其他領域技術與專利之用。

關鍵字

第六代行動通訊專利（6th Generation，6G） ； 文本挖掘 ； 主題建模 ； 隨機森林 ； 決策試驗與評估實驗室法(Decision Making Trial and Evaluation Laboratory，DEMATEL) ； 多準則決策分析 (multiple criteria decision making，MCDM)

並列摘要

The sixth-generation (6G) wireless communication is the foundation of the next-generation wireless communication. At present, the world is still in the exploratory stage, and mainstream technology is not yet clear. Thus, the current situation of patent development and landscape needs to be mined, where the results can serve as the basis for future academic researches, as well as foundations for developments of specifications and products. Albeit the mining of 6G patents is very important, related research is very little. Therefore, this research proposes a novel analytic framework which integrates text mining techniques and hybrid multiple criteria decision making (MCDM) methods. The analytic framework is used in mining patents of 6G.
Thus, the research first mined the patent database of the United States Patent and Trademark Office (USPTO). According to the 6G patents mined, topics were extracted by using the Latent Dirichlet Allocation (LDA). Then, for each topic, the Random Forest regression algorithm was introduced to obtain the feature importance of each topic of 6G technology. After that, the feature importance matrix was transferred into the initial impact matrix of Decision Making Trial and Evaluation Laboratory (DEMATEL). By adopting the DEMATEL based Analytic Network Process (DANP) method, the influence weight of each topic was obtained. In each topic, the technology with the highest weight is the driving force for future 6G wireless communication technology.
According to the results from empirical research, three key technologies were obtained: holographic radio technology, co-frequency co-time full-duplex, and wireless AI fusion. These techniques are the most important ones for the development standards and protocols of future 6G communication technology. And holographic radio technology is the most influential technology. The research results not only can serve as the basis for understanding the development of 6G technology, but can also provide a reference for the future development of 6G communication technology. In addition, the developed technical analysis framework can also be used to explore technologies and patents in other fields.

並列關鍵字

The Sixth Generation (6G) Communication ； text mining ； topic modeling ； Random Forest ； Decision Making Trial and Evaluation Laboratory (DEMATEL) ； multiple criteria decision making (MCDM)

參考文獻

Aazhang, B., Ahokangas, P., Alves, H., Alouini, M. S., Beek, J., Benn, H., ... & Chen, F. (2019). Key Drivers and Research Challenges for 6G Ubiquitous Wireless Intelligence (White Paper). Oulu, Finland: 6G Flagship, University of Oulu, 1.

Abraham, B. P., & Moitra, S. D. (2001). Innovation assessment through patent analysis. Technovation, 21(4), 245-252.

Alcacer, J., & Gittelman, M. (2006). Patent citations as a measure of knowledge flows: The influence of examiner citations. The Review of Economics and Statistics, 88(4), 774-779.

Al-Dulaimi, A., Wang, X., & Chih-Lin, I. (2018). 5G Networks: Fundamental Requirements, Enabling Technologies, and Operations Management. Hoboken, NJ: Wiley.

Allahyari, M., Pouriyeh, S., Assefi, M., Safaei, S., Trippe, E. D., Gutierrez, J. B., & Kochut, K. (2017). A brief survey of text mining: Classification, clustering and extraction techniques. arXiv preprint arXiv:1707.02919.