TechWatchTool: Innovation and Trend Monitoring

Proceedings of Recent Advances in Natural Language Processing, pages 660–665, Hissar, Bulgaria, 12-14 September 2011. TechWatchTool: Innovation and Trend Monitoring Hong Li Feiyu Xu German Research Center for Artificial Intelligence (DFKI), LT-Lab Alt-Moabit 91c, 10559 Berlin, Germany {lihong,feiyu,uszkoreit}@dfki.de http://www.dfki.de/lt/ Hans Uszkoreit Abstract In this paper we present an information service system that allows users to search for the key players of requested technol- ogy areas and for their collaboration net- works. This system utilizes information extraction and wrapper technologies for detecting persons, organizations, publica- tions and patents as well as relationships among them. Furthermore, it applies rela- tion extraction to detect statements on the web that indicate innovation trends. Vari- ous visualization methods are provided to let users monitor key players, their net- works and technology trends in a comfort- able way. 1 Introduction The innovation cycle of technologies is getting shorter and shorter. In recent years, many com- panies became aware of the potential of advanced information technologies for the efficient discov- ery and analysis of useful information in large vol- umes of online data such as business news, busi- ness reports, scientific publications and patents. Exploring patents or publications is an impor- tant approach to analyzing the trends of tech- nology development. Therefore, several systems emerged recently, which attempt to describe and predict the technology development trend based on the analysis of patents or publications (e.g., (Yoon and Park, 2004), illumin8 system1, Google Trends2(Rech, 2007), BlogPulse3 (Glance et al., 2004) and Collexis4). Most available systems are mainly based on a combination of statistical meth- ods and string match. There is still a big potential to apply language technologies to this task. 1http://www.illumin8.com/ 2http://www.google.de/trends 3http://www.blogpulse.com/ 4http://www.collexis.com/products In this paper, we present a system named TECH- WATCHTOOL5, that has already been success- fully tested by corporate users. In daily opera- tion, it now aids companies and analysts in de- tecting emergent technologies and in identifying associated key players, their cooperative networks and new trends that are relevant for their business sector. TECHWATCHTOOL applies methods from bibliometrics, information wrapping, information extraction and data mining. Language technology plays a central role in the extraction of names and technologies. The system monitors technologies with three modules: 1) a retrieval and extraction module for publications and patents for identifica- tion of key players and their relations, 2) a trend identification module and 3) an ontology-based navigation module. Furthermore, TECHWATCH- TOOL provides different views of the discovered data, which facilitate understanding and interpre- tation of the results. The remainder of the paper is organized as fol- lows: Section 2 explains existing systems for tech- nology and trend monitoring. Section 3 introduces the NLP tools used in TECHWATCHTOOL. Sec- tion 4 describes our system architecture and the core modules. Section 5 explains the result visual- ization and presentation. Finally, Section 6 gives a short conclusion. 2 Related Work Yoon and Park (2004) present a method to create patent networks with text mining methods to in- vestigate the technology development. Patents are represented as nodes in a …

Content truncated to 500 words