Changes between Version 1 and Version 2 of AutomaticTagAttribution


Ignore:
Timestamp:
12/04/09 10:11:34 (10 years ago)
Author:
kostas.christidis
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • AutomaticTagAttribution

    v1 v2  
    88 
    99 
    10 Automatic tag attribution is prone to error, it requires significant Natural Language Processing capability that is currently not on offer and the risk of the system assigning the wrong tags would significantly impair other functionality, particularly the significant retrieval functionality. There is scope within OrganiK to manually perfect the taxonomies used within each implementation. This is paraphrased the ‘Excel-question’. The Microsoft spreadsheet program Excel offers significant capabilities but is still user friendly to people that do not desire to utilise all the functions. The aim for OrganiK is to emulate a similar type of interface where maintenance to the taxonomy can be done where needed, as such improving the recommended tags. Improving the taxonomy significantly improves the results of the automated tag suggestions. However, these will not be attributed automatically as it requires the user to analyse the quality of the recommended tags.  
     10Automatic tag attribution is prone to error, it requires significant Natural Language Processing capability and the risk of the system assigning the wrong tags would significantly impair other functionality, particularly the significant retrieval functionality. There is scope within OrganiK to manually perfect the taxonomies used within each implementation. This is paraphrased the ‘Excel-question’. The Microsoft spreadsheet program Excel offers significant capabilities but is still user friendly to people that do not desire to utilise all the functions. The aim for OrganiK is to emulate a similar type of interface where maintenance to the taxonomy can be done where needed, as such improving the recommended tags. Improving the taxonomy significantly improves the results of the automated tag suggestions. However, these will not be attributed automatically as it requires the user to analyse the quality of the recommended tags.  
    1111 
    1212