以分類強化方法協助專利維護決策之制定

To Renew or Abandon My Patents:An Enhanced Classification Approach for Supporting Patent Maintenance Decisions

林蓓妤(Pei-Yu Lin)

指導教授 ： 魏志平

國立臺灣大學/管理學院/資訊管理學研究所/碩士(2015年)

https://doi.org/10.6342/NTU.2015.10249

摘要

專利已經成為企業最重要的智慧財產以及知識財產之一，透過法律在一定期間內所賦予的合法排他權利，發明人可以保護自己的創新發明不被他人所抄襲或利用，藉以提升競爭優勢。維護費的繳納可以視為發明人獲得這項保障的方法，但是一個企業或組織內部的專利相當龐大，維護費可能造成研發與製造上的巨額負擔，因此企業傾向維護他們認為較有價值的專利，將資源更有效地分配在有前景的研發上，此時維護決策就變得相當重要且迫切。
本研究預計使用過去研究中認為會影響專利維護的因素，以資料探勘的方式建立專利維護決策支援系統，並且透過分類強化的方法，來提升判斷的準確性。這種預測及支援系統可以協助企業與組織快速的判斷是否該對專利進行維護，增進決策的效率與準確性，同時降低決策成本。此外我們也利用美國專利局資料庫中取得的專利進行實證研究，結果顯示，我們所提出的專利維護決策支援模型表現優於基準模型(未引入分類強化方法之模型)。

關鍵字

資料探勘 ； 專利探勘 ； 專利智慧 ； 分類強化 ； 專利維護

並列摘要

Patents have become one of the most important intellectual properties and knowledge assets. With the legal exclusive rights provided by laws for a certain period, inventors can protect their inventions or products from being plagiarized or utilized, promote overall competitiveness, and gain a strategic advantage. Payments of maintenance fee can be regarded as the way inventors acquire such protection continually. However, the large portfolio of patents owned by large companies or organizations may cause heavy financial burden to them. Companies tend to maintain patents with greater value and allocate their resources to promisingtechnologies. In consequence, patent maintenance decisions are essential and urgent.
In this study, we examine factors influencing patent maintenance decisions which are considered significantby previous studies and propose a systematic solution to analyze patents for recommending patent maintenance decisions. Anenhanced classification algorithm is also applied in the study to improve the overall accuracy of our predictions. The patent maintenance decision support system can efficiently and effectivelyhelp companies and organizations determine whether a patent should be maintained while decreasing the decision making costs. In addition, we conduct experiments on the large scale United States Patent and Trademark Office (USPTO) database which contains over millions of granted patents. The empirical evaluation results show that our proposed model outperforms the benchmark model without the use of the enhanced classification algorithm.

並列關鍵字

Data Mining ； PatentMining ； PatentIntelligence ； Enhanced Classification ； Patent Maintenance ； Patent Renewal

參考文獻

Agrawal, R., & Srikant, R. (2001). On integtaing catalogs. Proceedings of the 10th International Conference on World Wide Web, 603–612.

Albert, M. B., & Avery, D. (1990). Direct validation of citation counts as indicators of industrially important patents. Research Policy, 20(3), 251–259.

Bessen, J. (2008). The value of U.S. patents by owner and patent characteristics. Research Policy, 37(5), 932–945.

Breiman, L. (1996). Bagging predictors. Machine Learning, 140(24), 123–140.

Chang, C.T. (2010). Supporting patent maintenance decision: A data mining approach. Unpublished Master Thesis, Institute of Service Science, National Tsing Hua University.