Learning new Term Weighting Schemes with Genetic Programming

Abstract : Text Classification (or Text Categorization) is a popular machine learning task which consists in assigning categories to documents. Feature weight methods are classic tools that are used in text categorization in order to assign a score to each term of a document based on a mathematical formula. In this paper, we are interested in automatically generating these formulas based on genetic programming. We experiment the generated formulas on three well-known benchmarks and state of the art classifiers.
Complete list of metadatas

Cited literature [21 references]  Display  Hide  Download

https://hal.inria.fr/hal-01662138
Contributor : Ahmad Mazyad <>
Submitted on : Tuesday, December 12, 2017 - 5:23:34 PM
Last modification on : Friday, February 2, 2018 - 1:02:17 AM

File

pg.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01662138, version 1

Collections

Citation

Ahmad Mazyad, Fabien Teytaud, Cyril Fonlupt. Learning new Term Weighting Schemes with Genetic Programming. The Biennial International Conference on Artificial Evolution, Oct 2017, Paris, France. 2017. ⟨hal-01662138⟩

Share

Metrics

Record views

34

Files downloads

31