Taxonomy Manager
One of the unique features of the Sysero System is the ability to combine searching with taxonomy browsing. However building a taxonomy can be an arduous task and the Taxonomy Manager is designed to be used by PSL's and KM staff to make this job considerably easier. As the taxonomy structure is held in a standard SQL database, exiting filing plans and taxonomies can be imported to get a quick start. We can also supply pre-built taxonomies based on the filing structures used by the Practical Law Company and Westlaw (with permission). However the Taxonomy Manager can be used to build a taxonomy from scratch without any programming skills. Each evening, the taxonomy is applied to the source documents using the node clues making maintenance a simple, techie free process.
Node Management
To
build or maintain the taxonomy structure, PSL's and KM staff can use
familiar right click functions even through their browser.
Nodes can be added, renamed or deleted using right clicks and
restructured using drag and drop. Security is maintained at
node level so PSL's can be responsible for specific nodes and their
children but not effect other nodes.
Clue Generation
Once a node has been created rules have to be built to decide which documents are classified by the node. The Taxonomy Manager allows users to add Clues (which are basically synonyms) for each node. During the nightly classification process, documents that meet the criteria are added to the node depending on the match score against the Clues. Whilst this is a simple process to understand and implement, the biggest issue comes from developing the Clues themselves. What terms should be used to identify whether a document should appear in a node? Using basic terms, such as the node name, the Taxonomy Manager can build a first cut of included documents. The subject matter expert can identify any documents in this rough list that are particularly good examples of documents that should be classified under the node. A suggest link next to the document uses conceptSearching's Related Topics feature to identify terms in the document that that appear more often in the document than in the rest of the index. The user can select the most appropriate of these terms and apply these to the node. Once the nightly classification has been completed the results of the updated rules can be seen.