Thesis

Text Categorization is an important study in the field of Text-Mining, with a wide range of applications. In recent years, through the development of Neural Networks, many techniques have been developed such as pre-trained language models, which are applicable to Natural Language Processing (NLP). Currently, the best practice for categorizing texts, e.g. writer recognition, is the application of Pre-Trained Language Models through Fine-Tuning. In this research, we analyze and present the application of the Universal Language Model Fine Tuning technique (ULMFiT) in some text categorization applications, which is developed by NLP's fast.ai research team. Furthermore, we compare this technique with others, and we conclude, presenting the results of this comparison.