TEXT MINING FOR PATENT MAP ANALYSIS

This paper describes some methods for patent map analysis based on text mining techniques and experiments on a realworld patent map created for an important technology domain: “carbon nanotube” show that most important category-specific terms occur in the machine-derived extracts from each patent segment.

Abstract

Patent documents contain important research results. However, they are lengthy and rich in technical and legal terminology such that it takes a lot of human efforts to analyze them. Automatic tools for assisting patent analysis are in great demand. This paper describes some methods for patent map analysis based on text mining techniques. We experiments on a realworld patent map created for an important technology domain: “carbon nanotube”. We show that most important category-specific terms occur in the machine-derived extracts from each patent segment and that machine-derived feature terms can be as good as those selected manually. The implications is that future patent mapping can be done in a more effective and efficient way.