# Algorithmes de classification et d'optimisation: participation du LIA/ADOC \'a DEFT'14

## TL;DR

This paper presents three statistical classification systems developed for the DEFT'14 campaign to identify the session of previous TALN conference articles, achieving a micro-precision of 0.76 through system fusion.

## Contribution

The paper introduces three novel statistical systems for session identification in conference articles and demonstrates the effectiveness of their fusion for improved accuracy.

## Key findings

- Micro-precision score of 0.76 on test corpus
- Fusion of systems enhances classification performance
- Effective approach for text classification in conference proceedings

## Abstract

This year, the DEFT campaign (D\'efi Fouilles de Textes) incorporates a task which aims at identifying the session in which articles of previous TALN conferences were presented. We describe the three statistical systems developed at LIA/ADOC for this task. A fusion of these systems enables us to obtain interesting results (micro-precision score of 0.76 measured on the test corpus)