Details of keyword search (sorry for the dutch content)
- bool: hond AND kat AND paard: a document is more interesting as more
of the different keywords occur.
- freq: hond OR kat OR paard: a document is more interesting as more
keywords occur.
- weighted: hond OF kat OF paard weighted: using the tf.idf
and normalization of document length.
For indexing and retrieval, we used the venerable SMART program, that combines
tf.idf word-weighting with the Vector Space Model.
|