When setting up a training set to be used for the creation of a conceptual index in Relativity, be sure to exclude computer generated files. Include conceptually rich documents in the set. Excel spreadsheets will rarely be examples of conceptually rich documents. Databases should also be avoided.
Conceptually rich documents should account for the full range of issues in a case and any new ones that arise.
Documents with foreign terms may be included if they are pertinent to key concepts.
Documents with lots of isolated strings of words should be avoided if possible.
Both documents which are relevant to a case and clearly non-relevant documents should be included. A review with conceptual analytics will want to find documents which should be excluded from a production.
Short emails of a few sentences are generally not good to use for a set of conceptually rich documents.
These operations can be run on a set of documents for which a conceptual index has been prepared:
3. Concept searching
4. Keyword expansion
5. Find similar documents