Combining the sample and feature selection method based on an artificial immunization algorithm for the patent quality classification model

• Main Authors: Hsuan-Ching Chen, 陳玄淨
• Other Authors: Pei-Chann Chang
• Format: Others
• Language: zh-TW
• Published: 2017
• Online Access: http://ndltd.ncl.edu.tw/handle/swe5zj

碩士 === 元智大學 === 資訊管理學系 === 105 === This study uses the Thomson Reuters' patent for the renewable energy in the Thomson Innovation patent database provided by Thomson Reuters. Using the combined sample and feature selection method and the patent quality classification model of artificial immune algorithm. The aim is to have a new patent at the beginning, we can classify whether the quality of this patent is of high value. This paper mainly uses the legal status of this patent quality standard to define the high, medium and low quality of this patent. Using the classification error rate to combine the sample selection and feature selection with the immune network of the artificial immune algorithm to establish the patent quality classification model, and finally by the experimental results can be found in this study proposed method compared to other methods of excellence.