wiki:UseCaseTaggingWithTextClassificiation

Version 3 (modified by heim, 18 years ago) (diff)

added some good usability stuff

implemented in TaggingApi, PimContentTraining

Prerequisite

A thing in the pimo is trained to some words. The project "branchoffice" is trained to the documents "businessplan" and "paul's address" and "Description of office building". The paul:!ProductEncyklopedia is trained to some garbage website.

Paul browses to a useful website "office supplies" related to branchoffice.

Result

website "office supplies" is tagged (occurrence-link) to thing "branchoffice".

Trigger

Paul clicks "tag this website" in Firefox.

Procedure

  • Paul browses to website "office supplies"
  • Paul clicks "tag this"
  • Website is transfered to TaggingApi, taggingapi checks fulltext using PimContentTraining
  • TaggingApi suggests "surely be branchoffice, may be ProducktEncyklopeid"
  • Paul clicks on "branchoffice"
  • tagging happens in TaggingApi

Alternatives

  • No thing found during search, the most (or the 5 most?) used tags should be displayed an highlighted (So that Paul kan start typing and the tag gets deleted automatically); Paul chooses to use this one or just enters the tag by himself, see use case: UseCaseTagging