PatNet: A Lexical Database for the Patent Domain

In the patent domain Boolean retrieval is particularly common. But despite the importance of Boolean retrieval, there is not much work in current research assisting patent experts in formulating such queries. Currently, these approaches are mostly limited to the usage of standard dictionaries, such as WordNet, to provide synonymous expansion terms. In this paper we present a new approach to support patent searchers in the query generation process. We extract a lexical database, which we call PatNet, from real query sessions of patent examiners of the United Patent and Trademark Office (USPTO). PatNet provides several types of synonym relations. Further, we apply several query term expansion strategies to improve the precision measures of PatNet in suggesting expansion terms. Experiments based on real query sessions of patent examiners show a drastic increase in precision, when considering support of the synonym relations, US patent classes, and word senses.