Recurrent Neural Networks - An introduction to binary classification using Natural Language Processing
Abstract
This essay outlines the applicability of Recurrent Neural Networks for supervised binary classification tasks in Natural Language Processing. It gives a general overview of the relevant theory and different recurrent layers. The theory is than applied to a binary classification task of offensive German language in twitter posts. Different model structures are compared to discover dependencies on batch size and number of iterations for generalization.
Type
Publication
In Learning deep - Perspectives on Deep Learning Algorithms and Artificial Intelligence
Felix Süttmann
Applied Statistician
Applied statistician with background in economics. I am interested in topics related to text analysis, NPL, Bayesian statistics and sustainability.