POSTECH Bio-Text Mining System Workbench


    Abstract

    The POSBIOTM/W is a workbench for machine-learning oriented biomedical text mining system. The POSBIOTM/W is intended to assist biologist in mining useful information efficiently from biomedical text resources. To do so, it provides a suit of tools for gathering, managing, analyzing and annotating texts.



  • System Description
    • The POSBIOTM/W workbench is intended to assist biologists in effectively and efficiently mining useful and valuable information from rich biomedical text resources.


    • The Managing tool helps biologists search, collect and manage literatures relevant to their interest. Moreover, the incremental PubMed search assistant provides some facilities for better utilization of PubMed.
    • The NER tool is a client tool of POSBIOTM/NER system and enables users to automatically and effectively annotate biomedical-related entities.
    • The Event Extraction Tool is a client tool of POSBIOTM/Event Extraction system and automatically extract interactions among four different entity types.
    • The GUI-based Annotation tool is designed to manipulate the manual annotations. The well-annotation data which is uploaded to the POSBIOTM system will contribute to the incremental build-up of a massive amount of named-entity and event annotation corpus.
  • Requirements
    • O/S : Windows 98, 2000, XP and Linux and Unix
      (all O/S should support Java runtime environment or Java web start)
    • Java Runtime Environment(JRE) v.1.5.0 or higher
      (it can be executed with lower version of JRE, but it can work improperly.)


  • Usage




    • Input the keyword of search to Quick Search Bar on the Search tab.
      • You can select number of abstract to be displayed by choosing "# per page".
      • Only Pubmed Search is available now.
      • If you click "MORE" button, you can execute Search Assistance, which will help you to find abstract by offering more query constraints.



    • Select one of the abstracts in the Search Result Frame.
      • If you select and double click one of the abstracts in Search Result Frame,
      • Text of the abstract will be shown in the Abstract Display frame.




    • If you want to do Named-entity Recognition task,
      • Click the NE tagging button on the Annotation tab
      • You will see the Named-entity recognized text.(it may take about 10 seconds)
      • Named entities in the text have different colors.
      • You can change the color of entity.
        And you can remove or add the entities with NE Annotator.
    • If you want to Event extraction task,
      • Click the Extraction button on the Annotation tab
      • You will see the event between the biological Named-entities.(it may take about 20~30 seconds)
      • You can add, remove and edit the Event with Event Annotator.



      • If you want to add a new event,
        • Click "Add" button in main window,
        • And select sentence number from "S-number" in Event Extraction Annotator frame.
        • And click the "Get Sentence" button.
        • Then you can see the sentence which is for adding event.
        • And specify interaction, effecter and reactant. Finally, click the "Confirm" button.
      • If you want to remove an event, select one of event you want to remove and click the "remove" button in the main window.
      • Update an event is similar to add a new event.


Top | Home | Back to the research page

 

San 31, Hyoja-Dong, Pohang, 790-784, Korea
Phone: +82 54 279 5581 | Fax: +82 54 279 2299