The study of the difference between the classification codes assigned to pre-grant publications and their corresponding issued patents

• Main Authors: Chan-Yi Lin, 林展逸
• Other Authors: Chung-Huei Kuan
• Format: Others
• Language: zh-TW
• Published: 2014
• Online Access: http://ndltd.ncl.edu.tw/handle/15973604152775177318

碩士 === 國立臺灣科技大學 === 專利研究所 === 102 === Most analyses are conducted using issued patents as subjects. However, these patents are granted after lengthy examination processes. Therefore, analytic results from the granted patents actually reflect the research activities exercised a long time ago. To reduce the above-described time delay, the only possible way is to use the so-called pre-grant publications that are open to public 18 months after the applications are filed. The pre-grant publications and their granted patents are both assigned classification symbols representing the technical fields to which the applications belong. Pre-grant publications therefore can be the subjects for the so-called patent classification analysis, which is a quite common and also important analytic method. The classification symbols assigned to pre-grant publications and their granted patent may very possibly be different. If the two are very different, using patent classification analysis on pre-grant publications to observe the research activities may be dubious. Yet if the differences are ignorable, the analytic result from using patent classification analysis on pre-grant publications may accurately reflect the research activities happening 18 months ago. This study compares the Current United States Patent Classification codes assigned to United States utility patents issued in 2012 and their pre-grant publications in three ways: (1) considering only the class code of primary classification; (2) considering only the primary classification code; and (3) considering all classification codes without discrimination. The study finds that only (1) considering only the class code of primary classification is more appropriate for patent classification analysis using pre-grant publications as there are 77.89% of the issued patents have the same class codes as their corresponding pre-grant publications. As to (2) considering only the primary classification code, the ratio drops significantly to 36.42%. As to (3) considering all classification codes, the study further classified the comparison results into five categories: (a) there are only 6.37% of the issued patents whose sets of classification codes are identical to those of the corresponding pre-grant publications, (b) there are 38.30% of the issued patents whose sets of classification codes are entirely different from those of the corresponding pre-grant publications, (c) there are 26.84% of the issued patents whose sets of classification codes contains those of the corresponding pre-grant publications, (d) there are 4.55%% of the issued patents whose sets of classification codes are contained in those of the corresponding pre-grant publications, and (e) there are 23.95% of the issued patents whose sets of classification codes have non-empty intersection with those of the corresponding pre-grant publications.