Recognizing Feeling in English Textual Exchanges Using BERT and BidLSTM Prototypes
Arvind Kumar Singh, Abhinandan Tripathi · 2024
One of the trickiest issues in automated language understanding is emotion recognition. Understanding human emotions through writing without seeing a person's face is thought to be a difficult undertaking. Thus, the machine learning community has been encouraged recently to construct a system that can distinguish between distinct emotions and understand the context of the utterances. We suggest a system that makes use of deep learning techniques to identify emotions. The AffectiveTweetsWeka package provided a set of psycholinguistic qualities, together with word embeddings derived from GloVe and BERT, for example) provide the primary input of system. The suggested system (EmoIden2) combines a BidLSTM neural network with a fully connected neural-network building design to produce performance results that significantly outperform the model of basline supplied through the Severin (2019) / Task:3 cordinator (Fig1-total 0.581, Fig1-total 0.748).