世界知識產權組織指出善加利用專利資訊，可縮短研發時間 60%以及節省研發經費 40%，所以完整的專利檢索加上詳盡的專利分析能釐清欲發展的技術領域的專利保護情形與阻力，進而對整個技術領域有個通盤的了解。
現今的專利檢索和分析須仰賴專利工程師以人工方式進行，此舉浪費大量人力物力，加上專利文件數快速成長，無疑是雪上加霜。
本論文模擬專利侵權比對中的專利範圍解讀，利用自然語言的頗析器將專利的申請範圍拆解成各項元件、元件的關係、步驟、步驟間的關係，並建立出機讀式檔案方便比對以及透過圖形化呈現此結構，最後提供步驟的相似度比對，提醒專利工程師哪些步驟是相似的以加快專利分析解讀的速度。
兩步驟的相似度比對演算法主要是將步驟分解成子特徵並匹配之，再依其所屬的特徵類別進行對應的相似度計算方法，最後以權重加總出兩步驟的相似度值。經由實驗驗證得知，本論文的步驟相似度比對方法在專利文件上確實比自然語言的BLEU方法有效。

World Intellectual Patent Organization (WIPO) indicates that realizing patent information well can reduce the 60% research time and save 40% research budgets. Therefore, sufficient patent retrieve and detailed patent analysis can clarify the patent protection situation and the obstacle, and then further help for the global understanding on the whole technical domain.
Now patent retrieve and analysis is worked by patent engineer manually, which costs a lot of efforts. As the rapid increasing of patent documents, the expense is more awful.
In this study we simulate the claim explanation, taking apart the claim into four parts: elements, the relations between elements, steps, and the relations between steps. Then, this information was used to build machine readable formats, which is easy for comparison or showing structure graphically. At last, we propose a way to compare the similarity of two steps. The similarity value show patent engineers which steps are similar and increase the efficiency on claim explanation.
The comparison algorithm splits the steps to many features, matches these features into different feature classes, then calculates the similarity value in each class by the corresponding calculating way. Finally sum the similarity value of all matched features with different weight to get the similarity value of two steps. The experimental results show that our similarity comparison way is better than BLEU at the steps comparison of patents.

[1] United States Patent Trademark Office http://www.uspto.gov/
[2] 陳省三，“專利檢索及其應用”， 國立清華大學專利暨技術移轉培訓講義，新竹，2006
[3] 陳裕禎，“如何解讀專利範圍？”，國立清華大學專利實務暨技轉培訓班講義，新竹，2005
[4] Section 101 of Title 35 U.S.C.
[5] 陳志超，“專利法-理論與實務”，五南圖書出版股份有限公司，2002
[6] 林士能，“專利文件語意之擷取與比對”，國立清華大學資訊工程學系，碩士論文，2005
[7] Chi-Feng Lee, “Automatic Acquisition of Domain Specific Regular Expressions from Patent Documents, Master thesis, NTHU, 2006
[8] Kristina Toutanova and Christopher D. Manning, “Enriching the Knowledge Sources Used in a Maximum Entropy Part-of-Speech Tagger”, Proceedings of the Joint SIGDAT Conference on Empirical Methods in Natural Language Processing and Very Large Corpora (EMNLP/VLC-2000), Hong Kong.
[9] Dan Klein and Christopher D. Manning, “Accurate Unlexicalized Parsing”, Proceedings of the 41st Meeting of the Association for Computational Linguistics, 2003
[10] Marie-Catherine de Marneffe, Bill MacCartney, and Christopher D. Manning, “Generating Typed Dependency Parses from Phrase Structure Parses”, In Proceedings of the LREC Conference. Genoa, Italy. 2006
[11] W.S. Torgerson, “Multidimensional scaling of similarity”,Psychometrika,vol. 30, pp.379-393, 1965
[12] A. Tversky, “Deatures of similarity”,Psychological Review, vol. 84, pp. 327-352, 1977
[13] A.B. Markman and D. Gentner, “Structural alignment during similarity comparisons”, Cognitive Psychology, vol. 25, pp.431-467, 1993
[14] George A. Miller, “WordNet: A Lexical Database for English”, Communications of the ACM, Vol.38 No.11, 1995
[15] 王世仁，專利工程導論，俊傑書局，2002
[16] Holland, J., 1975, “Adaptation in Natural and Artificial Systems”, University of Michigan Press, Ann Arbor, MI.
[17] Kishore Papineni, Salim Roukos, Todd Ward, and Wei-Jing Zhu, “BLEU: a Method for Automatic Evaluation of Machine Translation”,Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics (ACL), Philadelphia, July 2002, pp. 311-318.