Text and Document Analytics for Legal Document Reviews Using Deep Learning

Convolution Neural Networks provide effective means of text classification. Hitherto research in text is performed on the short documents containing passages and small texts. The challenge of large legal text documents has been carried out in this experiment using a 2-CNN model, which is a vanilla model performant as a multi-class classification. Datasets are collected from the local legal authorities which are already digitized and converted to pure text documents. The results shown in the research support the legal experts to use machine learning for the multi-class classification of the large legal text documents to identify the classes with the key sentences/phrases.