Analysis of Patents for Prior Art Candidate Search
Résumé
In this paper, we describe a method for analyzing a collection of patents in order to help prior art candidate search in an interactive and graphical way. The method relies on the use of two data mining methods: hierarchical agglomerative clustering and principal component analysis, which are applied successively. The correlation between the application patent and the other patents is a good indicator to help decide the classes of patents to look at.
| Origine | Fichiers produits par l'(les) auteur(s) |
|---|