Co-word occurrence network analysis in patents

Co-word occurrence network is a network in which nodes represent words and edges denote their joint appearance in a paper.

Given below is a co-word occurrence network of words in Patent titles from patent records downloaded from U.S. Patent and Trademark Office patents (USPTO) database on 1000 patent records related to word ‘Computer’ were downloaded.

The .csv file is preprocessed- the contents in the field Title are normalized and Word Co-Occurrence Network extracted. The network has 999 word nodes and 8144 co-occurrence edges with no isolated node(s).

To keep only the strongest edges in the network we extract a network with top 1000 edges by weight and visualize it using GUESS. The network is shown below. Only the co-word occurances with more than 40 reference have been labeled as shown below:


The network shows the frequently occurring words in the patent title of the patents data related to term‘computer’.

Leave a Reply

Your email address will not be published. Required fields are marked *