Can Domain-Adaption improve the performance of Sentiment Analysis?
Does Domain Adaption have any effect of results in Sentiment Analysis? I am going to train a BERT language model based on some texts particularly in Health area, then I want to apply Opinion Mining on that to find which text carries positive or negative sentiment. I have run it on pre-trained BERT and get some results, however my question is that, does Domain Adaption help me to increase the performance of my model?
Topic bert sentiment-analysis language-model
Category Data Science