Μοντέλα transformer για ανάλυση συναισθημάτων
Transformer models for sentiment analysis
View/
Keywords
NLPAbstract
The study and processing of natural language (NLP) is a field of
study that has shown significant progress during the last years.
During this time period there have been various models that have
been invented, applied and evaluated that have contributed to the
overall progress of this field of study.
The contemporary approach to the subject of natural language
processing is the usage of machine learning and specifically the
usage of a neural network architecture called transformer model.
This deep learning architecture has inherited various
characteristics from previous approaches and is currently capable
of covering a vast field of natural language processing tasks.
One of these tasks is called Sentiment Analysis. There have been
many models trained for this specific task with the end goal of
determining whether the sentiment of the text is overall positive or
negative. Sentiment analysis is a field that bridges linguistics with
computer science and has as an end goal the classification of the
text to a specific sentiment either positive or negative.
Sentiment analysis is a useful and versatile tool that is being
utilized by most social media analysts due to its ability to
determine the overall sentiment of a text by taking into account
the meaning, the words and the structure of the text between other
factors.