Document Indexing

1. Goal = Find the important meanings and create an internal 

              representation

2. Factors to consider:

    Accuracy to represent meanings (semantics)

        Exhaustiveness (cover all the contents)

        Facility for computer to manipulate. 

3. What is the best representation of contents?

        Char. string (char trigrams) : not precise enough 

        Word: good coverage, not precise

        Phrase: poor coverage, more precise

        Concept: poor coverage, precise

Coverage (Recall)Accuracy (Precision)

                                    String      Word       Phrase      Concept