![]() |
|
Approach
In the first phase of the project available technologies for the indexing and information extraction will be used to automatically index a subset of the documents of FIZ Technik and TIB Hannover. The results will be evaluated and compared with the manually generated index. In the next step, the different techniques will be combined in order to exploit the benefits of the different approaches.
The semi-automatic indexing will support the experts in their job by proposing appropriate index terms, which can afterwards be corrected. With the help of self-learning algorithm the index algorithm should learn over time to improve its suggestions.
The results of the indexing process also support the users of the system. The user should be assisted in the query formulation by proposing related terms. After the query execution, the result set contains not only the description of the documents, but also related index terms, which can be used to further refine the query. Finally the retrieval engine, which handles the user queries, combines linguistic methods and thesauri to improve the query results.
After three years
At the end of the project FIZ Technik and TIB Hannover will be able to index new documents much more efficient and with a higher quality. Thus strengthen their position in the global information market, as their customers will also get high quality and up to date information.
Project Partners:
- Fachinformation Technik (FIZ Technik), Frankfurt
- IAI - Institut der Gesellschaft zur Förderung der Angewandten Informationsforschung e.V., Saarbrücken
- Technische Informationsbibliothek Hannover (TIB Hannover)
- L3S Research Center, Hannover



Research Areas
