Social relations play an important role in a real community. Interaction patterns reveal relations among actors (such as persons, groups, firms), which can be merged to produce valuable information as a network structure. This paper presents a new approach to extract inter-firm networks from the web for further analysis. Extraction of relations between a pair of firms is realized using a search engine and text processing. Because names of firms co-appear coincidentally on the web, we propose an

more »

... advanced algorithm, which is characterized by addition of keywords (relation keywords) to a query. The relation keywords are obtained from the web using a Jaccard coefficient. As an example, we extract a network of 60 firms in Japan including IT, communication, broadcasting, and electronics firms from the Web and show comprehensive evaluations of our approach. Our method is a first attempt to extract inter-firm networks from the web using a search engine. The approach is also applicable to other actors, such as famous persons, organizations or other multiple relational entities.