Table 1.
Overview of literature review.
Fig 1.
Aspect-based sentiment analysis process.
Table 2.
Statistics of cricket dataset.
Table 3.
Statistics of restaurant dataset.
Fig 2.
BERT Encoder architecture.
Fig 3.
Random forest architecture.
Table 4.
Hyperparameters for tRF-BERT.
Table 5.
Hyperparameters for BERT.
Table 6.
Hyperparameters for RoBERTa.
Fig 4.
tRF-BERT Architecture.
Table 7.
Aspect detection for cricket dataset.
Table 8.
Sentiment classification for cricket dataset.
Table 9.
Aspect detection for restaurant dataset.
Table 10.
Sentiment classification for restaurant dataset.
Table 11.
Runtime comparison.
Table 12.
Accuracy comparison with different tokenizer.
Table 13.
Accuracy of different models while experimenting for building hybrid model.
Table 14.
Evaluating this study in relation to prior research on aspect detection in Bengali ABSA.