none

• Main Authors: Nian-Ze Wang, 王念澤
• Other Authors: Chung-Huei, Kuan
• Format: Others
• Language: zh-TW
• Published: 2018
• Online Access: http://ndltd.ncl.edu.tw/handle/374yd2

碩士 === 國立臺灣科技大學 === 專利研究所 === 106 === A patent classification is a system for examiners to categorize patent documents by its technical fields. We know that patent classification is extensively used to evaluate a company, calculate the strength of a patent and ,by inspecting the filed patent applications and documents, even the history of a specific technology. We may use these data to create a technology-function Matrix, which will help us to find some potential direction for research in the future. In 2013, Europe and America started using cooperative patent classification (CPC), enabling the examiners to give patent classifications more clearly by its technical field which has been very helpful for relevant researches of patent classification analysis. In previous researches, in order to analyse the patents at interest, extracting out the classification numbers and coming out with a statistic result is required. So that knowing which one(s) of the classifications are much more than the others. Moreover, utilizing the analysis enables us to determine the technical fields of these patents. Furthermore, analysis can also be done with the combination of other elements, such as countries, patentees, and etc. However, when a patent is given many classifications (by the examiners), despite that it means the patent could have covered a more extensive scope of art, it may indicate otherwise. It could have been that the requirement of two or more classifications are necessary to express the information about its technique more profoundly, instead of being understandable from one classification. As a result, we chose to case study about CRISPR, a popular and rapid-developing biological technique and analyze the related patents from the United State Patent and Trademark Office database. We want to find out, when a technique is developing, whether co-occurrence of patent classifications occurs, compare to the traditional patent classification analysis. In addition to looking at the results of the patents as a whole, there are also annual observations. We observe the number of classification group variation from year to year, and it does present some clues demonstrating the directions of future research. Finally, according to results of the research, we define some types of patent classification co-occurrence, for example, “dependent co-occurrence” and what match the theory of co-occurrence in the research. Furthermore, focusing on the patents corresponding to one classification group and manually distributing them, we verify them having the same technical field. As a result, the classification co-occurrence done by the research could be another way to filter, which could be helpful for reducing the time to search for patents.