Table 3 Comparative analysis of various sentiment classifier models in multiple domains
From: Analysis of customer reviews with an improved VADER lexicon classifier
Methods/Domain | Electronics | DVD | Kitchen | Books | Average |
|---|---|---|---|---|---|
SentiDraw [33] | |||||
A (%) | 80.9 | 78 | 78.1 | 78.4 | 78.85 |
P (%) | 60 | 63 | 65 | 58 | 61.5 |
R (%) | 62 | 65 | 70 | 72 | 67.25 |
F (%) | 66 | 68 | 58 | 55 | 61.75 |
SL-MLP [34] | |||||
A (%) | 77.6 | 72.4 | 76.3 | 73 | 74.825 |
P (%) | 55 | 58 | 66 | 81 | 65 |
R (%) | 42 | 44 | 46 | 48 | 45 |
F (%) | 45 | 48 | 53 | 58 | 51 |
CAN [40] | |||||
A (%) | 87.13 | 87.23 | 88.8 | 89.8 | 88.24 |
P (%) | 75 | 65 | 72 | 70 | 70.5 |
R (%) | 78 | 75 | 76 | 68 | 74.25 |
F (%) | 73 | 75 | 77 | 80 | 76.25 |
UDA-SP[41] | |||||
A (%) | 85.7 | 84.2 | 88.6 | 85.3 | 85.7 |
P (%) | 85 | 88 | 82 | 86 | 85 |
R (%) | 81 | 83 | 84 | 86 | 81 |
F (%) | 80 | 83 | 85 | 86 | 80 |
CDSARFE [42] | |||||
A (%) | 86.7 | 76.29 | 78.01 | 74.9 | 86.7 |
P (%) | 83 | 86 | 87 | 81 | 83 |
R (%) | 80 | 87 | 85 | 81 | 80 |
F (%) | 81 | 82 | 84 | 86 | 81 |
IVADER (proposed) | |||||
A (%) | 95.7 | 98.64 | 96.55 | 97.41 | 98.64 |
P (%) | 97 | 93 | 96 | 91 | 97 |
R (%) | 91 | 94 | 92 | 93 | 94 |
F (%) | 93 | 92 | 90 | 91 | 92 |