Computer assisted patent inventiveness evaluation and search system

• Main Authors: Yeh, Tzu-Liang., 葉子諒
• Other Authors: SHIU, DUNG-LIANG
• Format: Others
• Language: zh-TW
• Published: 2017
• Online Access: http://ndltd.ncl.edu.tw/handle/en4qa9

碩士 === 國立清華大學 === 工業工程與工程管理學系所 === 105 === This research modified the 5 levels of inventions presented by Altsuller and built an objective process and interpretations for classification of patent levels of inventiveness. Reviewing patents to determine its level of invention is a time-consuming task as it requires determining its level of novelty and comparison of many patents in its related application fields. This research developed a computer assisted software to assess level of invention based on comparison of novelty of action principles, novelty of featured functions in its application fields, and newness of its featured parameter set. The system developed by this research helps to evaluate patent inventiveness by using web crawler, document feature selection, similarity analysis. The overall contributions of this study include: 1) Providing objective inventiveness classification between level 2 and level 3 compared to the ambiguous classification in the original Altsuller's classification system. 2) The computer assisted system established enables users to more comprehensively compare various patents in the same application fields compared to manual patent searches. The system can be used to assess the novelty level of any idea as long as its concepts can be expressed in the six key elements of featured functions even if a patent application has not been filed. 3) Expanding the function/attribute lists with Wordnet synonyms enabling users to more comprehensively identify distinctive functions and attributes for patent assessments. 4) Classifying more than 900 scientific principles in accordance with the TRIZ field classification to facilitate more efficient novelty classification on the action principles of the subject patent idea.