專利訴訟常發生於擁有高品質專利或在產品技術重疊性高的企業之間。擁有高品質專利之企業，在專利訴訟上常扮演主動訴訟之角色。反之，專利相對劣勢之企業則大多處於被告之位置。本文收集近幾年之訴訟新聞資料後，以灰關聯分析找出在專利訴訟上互動關聯較強之企業，為主要專利訴訟玩家群（常為原告及被告者）。本文發現近幾年之LED專利訴訟之主要玩家分別為Nichia（日亞化學）、Everlight（億光電子）、Seoul（首爾半導體）、及Epistar（晶元光電）。而主要玩家訴訟之專利標的代表該專利在產業中的重要性，也顯示訴訟企業彼此的技術能力差異。故透過專利訴訟可確定企業在相關產業技術中之專利定位及角色。最後，本文利用機器學習方法（k-均值分群演算法與類神經網路）提出一預測模型，以協助企業性更系統性地判斷專利在未來可能之訴訟角色。該預測模型可減少大量人力判斷之過程，且協助企業有更一致性的決策品質。

Patent litigation usually occurs in enterprises that have high quality patents or have high overlap in the product technology. Enterprises with high-quality patents often play the role of active litigation in patent litigation. In addition, enterprises with weak patents mostly in the position of defendants. After collecting news materials of litigation in recent years, we used grey relational analysis (GRA) to find enterprises with strong interaction in patent litigation, which are the key players of patent litigation (there are usually in the role of plaintiff and defendant). This thesis finds that the key players in patent litigation of LEDs in recent years are Nichia, Everlight, Seoul and Epistar. The objects of patent litigation of key players are often the important patents in the industry, which shows the differences in the technological capabilities of the litigation companies. Therefore, through patent litigation, the company's patent positions in related industrial technologies can be determined. Finally, we used machine learning methods including k-means clustering algorithm and neural network to propose a predictive model to comprehensively predict the litigation role of patents in the future. The proposed model can reduce a large number of human judgments and help companies have more consistent decision-making quality.

[1] D. Yang, “Patent Litigation Strategy and Its Effects on the Firm,” International Journal of Management Reviews, vol. 21, no. 4, pp.427-446, 2019, doi: 10.1111/ijmr.12202.
[2] Y. G. Lee, "What affects a patent’s value? An analysis of variables that affect technological, direct economic, and indirect economic value: An exploratory conceptual approach," Scientometrics, vol. 79, no. 3, pp. 623-633, 2009, doi: 10.1007/s11192-007-2020-5.
[3] J. L. Deng, "Control problems of grey systems," Systems & control letters, vol. 1, no. 5, pp. 288-294, 1982, doi: 10.1016/S0167-6911(82)80025-X.
[4] W. W. Hines, D. C. Montgomery, and D. M. G. C. M. Borror, Probability and statistics in engineering, New York, NY, USA: John Wiley & Sons, 2008.
[5] J.L. Deng, The Theory and Application of Grey System, Taipei: Gau-Lih Book Co., 2000.
[6] X. Wang and H. Zhang, “August. Patent Management Evaluation of Tobacco Enterprises Based on Grey Relational Analysis and Efficacy Coefficient Method,” 2020 IEEE 8th International Conference on Information, Communication and Networks (ICICN), IEEE, pp. 188-192, 2020, doi: 10.1109/ICICN51133.2020.9205066.
[7] L. Xu and Y. Zhang, “Quality improvement of smart senior care service platform in China based on grey relational analysis and Fuzzy-QFD,” Grey Systems: Theory and Application, 2021, doi: 10.1108/GS-05-2020-0068.
[8] K.L. Wen, Y.L. Huang, W.C. Chang, T.C. Chang and C.J. Lai, Grey relational model method and application, Taipei: Gau-Lih Book Co., 2003.
[9] H. Kim and J. Song, "Social network analysis of patent infringement lawsuits," Technological Forecasting and Social Change, vol. 80, no. 5, pp. 944-955, 2013, doi: 10.1016/j.techfore.2012.10.014.
[10] W. L. Yang, “Exploring Roles & Positions of LED Companies Involved in Patent Infringement Litigations—A Social Network Analysis,” Technology Law Review, vol. 13, no. 2, pp. 149-186, 2016, doi: 10.3966/181130952016121302004.
[11] S. P. Lloyd, "Least square quantization in PCM. Bell Telephone Laboratories Paper. Published in journal much later: Lloyd, SP: Least squares quantization in PCM," IEEE Transactions on Information Theory, vol. 28, no. 2, pp. 129-139, 1982, doi:10.1109/TIT.1982.1056489.
[12] J.A. Hartigan, Clustering algorithms, New York, NY, USA: John Wiley & Sons, Inc. 1975.
[13] M. A. Wong, and J. A. Hartigan, "Algorithm AS 136: A k-Means Clustering Algorithm. Journal of the Royal Statistical Society," Series C (Applied Statistics), vol. 28, no. 1, pp. 100–108, 1979.
[14] R. Liang, Y. Li, X. Chen and J. Chen, August, “Patent Trend Analysis through Text Clustering based on K-Means Algorithm,” 2020 International Symposium on Computer Engineering and Intelligent Communications (ISCEIC), IEEE, pp. 115-118, 2020, doi: 10.1109/ISCEIC51027.2020.00032.
[15] N. Mittas and D. Mehrotra, “Clustering the Patent Data Using K-Means Approach,” Software Engineering, Springer, Singapore, pp. 639-645, 2019, doi: 10.1007/978-981-10-8848-3_61.
[16] D. J. MacKay and D. J. Mac Kay, Chapter 20: An Example Inference Task: Clustering, Information theory, inference and learning algorithms, Cambridge, CBG, UK: Cambridge university press, 2003.
[17] F. Rosenblatt, Principles of Neurodynamics: Perceptrons and the Theory of Brain Mechanisms, New York, NY, USA: Spartan Books, 1962.
[18] H. S. Stern, "Neural networks in applied statistics," Technometrics, vol. 38, no. 3, pp. 205-220, 1996, doi: 10.1080/00401706.1996.10484497.
[19] M. Lee, “An analysis of the effects of artificial intelligence on electric vehicle technology innovation using patent data,” World Patent Information, vol. 63, 2020, doi: 10.1016/j.wpi.2020.102002.
[20] R. Setchi, I. Spasić, J. Morgan, C. Harrison, and R. Corken, “Artificial intelligence for patent prior art searching,” World Patent Information, vol. 64, 2021, doi: 10.1016/j.wpi.2021.102021.
[21] M. Saade, M. Jneid, and I. Saleh, “Enhancing Decision-Making in New Product Development: Forecasting technologies revenues using a Multidimensional Neural Network,” European, Mediterranean, and Middle Eastern Conference on Information Systems 2020 Nov 25, Springer, Cham, pp. 715-729, 2020, doi: 10.1007/978-3-030-63396-7_48.