Mining foreign language information resources

Summary form only given. The availability of electronic information sources, such as publication and patent abstract databases, is a potent intelligence resource. Considerable work is underway to mine such resources. Such "information exploitation" usually is constrained to one's own language. Access to international R&D through visitation and conferencing is challenging and costly. Americans tend to be less proficient in multiple languages than, say, their European competitors. These factors prompt us to initiate efforts to mine foreign language publication and patent information resources. Taking advantage of abstract databases through text mining software, augmented by selective machine translation, could bolster international technological intelligence. This paper presents our initial foray into extracting information from non-English resources. Our evaluation applies software called the "Technology Opportunities Analysis System" (TOAS), which is oriented explicitly to technology management purposes.