IRISA at SMM4H 2018: Neural Network and Bagging for Tweet Classification

Anne-Lyse Minard (1) , Christian Raymond (1) , Vincent Claveau (1)

1 LinkMedia - Creating and exploiting explicit links between multimedia fragments

This paper describes the systems developed by IRISA to participate to the four tasks of the SMM4H 2018 challenge. For these tweet classification tasks, we adopt a common approach based on recurrent neural networks (BiLSTM). Our main contributions are the use of certain features, the use of Bagging in order to deal with unbalanced datasets, and on the automatic selection of difficult examples. These techniques allow us to reach 91.4, 46.5, 47.8, 85.0 as F1-scores for Tasks 1 to 4.

Domaines: Recherche d'information [cs.IR]; Informatique et langage [cs.CL]; Intelligence artificielle [cs.AI]

Anne-Lyse Minard, Christian Raymond, Vincent Claveau. IRISA at SMM4H 2018: Neural Network and Bagging for Tweet Classification. SMM4H 2018 - Social Media Mining for Health Applications, Workshop of EMNLP, Oct 2018, Brussels, Belgium. pp.1-2. ⟨hal-01937019⟩