Deep Learning for French Legal Data Categorization

Abstract

In current years, deep learning has showed promising results when used in the field of natural language processing (NLP). Neural Networks (NNs) such as convolutional neural network (CNN) and recurrent neural network (RNN) have been utilized for different NLP tasks like information retrieval, sentiment analysis and document classification. In this paper, we explore the use of NNs-based method for legal text classification. In our case, the results show that NN models with a fixed input length outperforms baseline methods.