Reexamining tf.idf based information retrieval with genetic programming
Nir Oren · 2002
The tf.idf family of vector based information retrieval schemes is very popular due to its simplicity and robustness, as well as its tractability to enhancements. This paper proposes a method to automatically perform a search for new tf.idf like schemes using genetic programming. The results of this automated search are then evaluated in a simple usage scenario. Also evaluated are the effects of using different fitness functions in the genetic programming phase.