A New Feature-Centered Method for Classifying Emotions Based on Context

Automated sentiment evaluation techniques are in great demand because of the rising tide of reviews on the internet as well as additional text with emotional nuance. Automatic, granular collection of emotion from written words or phrases is made possible by aspect-based categorization of sentiment (ABSC) owing to the complex structure of natural tongues, the study industry continues to encounter difficulties in accurately identifying and extracting them. In order to anticipate or categorize reviewers' emotions, the authors of this research provide a technique called Attention-based Gated Recurrent Neural Network (A-GRU). It presents the hybrid strategy by integrating word incorporation, POS tagging, & positional contextual data. In contrast to previous research, which has focused on single-channel Recurrent Neural Networks (RNNs), we propose a Multi-Channel Gated Recurrent Neural Networks (MC-GRU). Word insertion, POS tags, & context location information all work together to boost the precision with which aspects & their related attitudes may be identified & predicted. As a result of the filtration applied by the system of attention which initially identified the key phrases, this in turn aids in determining the attributes of entities connected to the stated attitude. Using industry-standard datasets, the empirical study demonstrates the efficacy of the suggested strategy in comparison to the state-of-the-art methods in the aforementioned literature. The experimental findings show that the suggested model outperforms the state-of-the-art methods for the F1-measure, having a total score of 94% on the component extraction test & 98% on the emotional categorization job.