Text classification is a common NLP problem that involves building a model trained on supervised text and distribute labels to unseen text data. Deep learning have proven to give state of the art performance for large datasets of images and speech, motivated by that in this paper we study application of deep learning on a custom small text dataset by testing them using different architectures and embedding types.
Application of deep learning techniques for text classification on small datasets
IJESC, International Journal of Engineering Science and Computing, Vol.8, N°4, April 2018
PERMALINK : https://www.eurecom.fr/publication/5559