A two-step agglomerative hierarchical clustering method for patent time-dependent data

Publication Type:
Conference Proceeding
Advances in Intelligent Systems and Computing, 2014, 213 pp. 111 - 121
Issue Date:
Filename Description Size
ISKE 2012 Hongshu Chen.pdf Published version446.4 kB
Adobe PDF
Full metadata record
© Springer-Verlag Berlin Heidelberg 2014. Patent data have time-dependent property and also semantic attributes. Technology clustering based on patent time-dependent data processed by trend analysis has been used to help technology relationship identification. However, the raw patent data carry more features than processed data. This paper aims to develop a new methodology to cluster patent frequency data based on its timerelated properties. To handle time-dependent attributes of patent data, this study first compares it with typical time series data to propose preferable similarity measurement approach. It then presents a two-step agglomerative hierarchical technology clustering method to cluster original patent time-dependent data directly. Finally, a case study using communication-related patents is given to illustrate the clustering method.
Please use this identifier to cite or link to this item: