|
Author | Year | Contribution | Results | Research gap |
|
Morio et al. [29] | 2020 | The author used PoS with CNN | F1: 51.55, P: 56.54, and R: 47.37 | More feature engineering can improve the performance. |
Jurkiewicz et al. [30] | 2020 | The authors used CRF | F1: 49.15, P: 59.95, and R: 41.65 | More data may improve performance. |
Chernyavskiy et al. [31] | 2020 | The authors used embeddings with CRF | F1: 49.10, P: 53.23, and R: 45.56 | Non-English language can be used. |
Khosla et al. [32] | 2020 | The authors used Bag of Words with LSTM | F1: 47.66, P: 50.97, and R: 44.76 | TF/IDF can be used with LSTM. |
Paraschiv and Cercel [33] | 2020 | The authors used embeddings with LSTM | F1: 46.6, P: 58.61, and R: 37.94 | More data can improve performance. |
Dimov et al. [34] | 2020 | The authors used n-Grams with LSTM | F1: 44.68, P: 55.62, and R: 37.34 | The authors have only used n-Gram feature other features may be considered in the future. |
Blaschke et al. [35] | 2020 | The authors used PoS with SVM | F1: 43.86, P: 42.16, and R: 45.7 | Different kernels of SVM can be explored in order to get more accuracy. |
Verma et al. [36] | 2020 | The authors used ELMo with CNN | F1: 43.60, P: 49.86, and R: 38.74 | More feature engineering is possible. |
Singh et al. [37] | 2020 | The authors used PoS with BERT | F1: 42.21, P: 46.52, and R: 38.63 | More data are required for BERT, and the authors only used 536 articles. |
Ermurachi and Gifu [38] | 2020 | The authors used Bag of Words with MNB and with LR | F1: 33.21, P: 24.49, and R: 51.57 | Other features like sentimental, and emphatic may improve the performance. |
Dewantara et al. [39] | 2020 | The authors used embeddings with CNN | F1: 23.47, P: 22.63, and R: 24.38 | More embeddings can be used in order to achieve better results. |
Daval-Frerot and Yannick [40] | 2020 | The authors used embeddings with RF | F1: 18.18, P: 34.14, and R: 12.39 | Bag of Words and Tf/IDF may be used with RF. |
|