TECHNOLOGY CLUSTERING OF PATENTS BASED ON THEIR PATENT CLASSIFICATION
Date Issued
2006
Date
2006
Author(s)
Chou, Yung-Ming
DOI
zh-TW
Abstract
In this thesis, patents are clustered into technological classifications according to their technological similarities that can be represented by the US patent class numbers. Ward’s method of hierarchical agglomerative clustering analysis is adopted to find hierarchical relationships in the clustering results, and the difference between patent class number features and the weight of patent class number features are studied as well. This research can be divided into three steps: the related patent class numbers are first rearranged through the patent classification system; secondly the technological similarities among all patents are formulated and computed; finally patents with higher technological similarity are clustered in accordance with the clustering analysis of patents.
Last but not the least, patents in ODF technology is proceed through our clustering analysis, and the practicality of applying the patent classification clustering in the classification of different patent technologies is verified. Moreover, a technological clustering result with hierarchical relation is established that fits what a researcher requires better than the existing technological professional classification did.
Subjects
相似性
華德法
叢集分析
similarity
clustering analysi
Type
thesis
File(s)![Thumbnail Image]()
Loading...
Name
ntu-95-R93522631-1.pdf
Size
23.53 KB
Format
Adobe PDF
Checksum
(MD5):f88e46c634ecc881185bc5148123cffe