Text Classification in the Legal Domain

Data analysis of legal texts, especially such that requires high precision, is not achieved over a few months of using random algorithms and relying on freely floating datasets with semi-trustable annotation quality. It requires a team of highly motivated people (preferably with diverse skills), high-quality datasets and a lot of time for experiments, failures, refactoring, testing, and development as well as heated discussions.In this talk he will present how Seal have worked with Machine Learning, Natural Language Processing, and data analysis of legal texts and the difficulties they have encountered and the approach they have taken to address them.