Open Access

Sentiment Analysis of Korean Modern Novel Texts Applying Deep Learning Models

  
Mar 26, 2025

Cite
Download Cover

Figure 1.

Basic flow chart of emotion analysis
Basic flow chart of emotion analysis

Figure 2.

The graphic structure of the chain CRF
The graphic structure of the chain CRF

Figure 3.

CBOW model structure
CBOW model structure

Figure 4.

BGRmodel
BGRmodel

Figure 5.

The bert-bgru-att model structure
The bert-bgru-att model structure

Figure 6.

Distribution of sentiment words in SST-2
Distribution of sentiment words in SST-2

Figure 7.

Distribution of sentiment words in IMDB
Distribution of sentiment words in IMDB

Figure 8.

“Mom’s Stakes” sets the emotional curve
“Mom’s Stakes” sets the emotional curve

Figure 9.

“Mom’s Stakes” grows emotional curves
“Mom’s Stakes” grows emotional curves

Results of the experiment on SST-2

Model serial number Model Accuracy rate (P) Recall rate (R) F1 value (F1-score)
(1) BERT 87.289 95.586 91.137
(2) BG 90.649 93.275 91.914
(3) BL 87.533 94.926 91.071
(4) BGA 91.323 92.212 91.748
(5) BLA 91.533 92.432 91.931
(6) BH 89.523 93.826 91.595
(7) BHW 90.595 93.092 91.816
(8) B_LEX 88.354 94.339 91.237
(9) BHW_LEX 88.24 95.146 91.557
(10) BERT-BGRU 89.935 94.082 91.976

Sst-2 data set

Sentence number Positive Negativity
Training set 6524 3472 3241
Verification set 785 393 382
Test set 1745 886 893

Results of the experiment on IMDB

Model serial number Model Accuracy rate (P) Recall rate (R) F1 value (F1-score)
(1) BERT 86.592 95.182 90.316
(2) BG 89.986 92.395 91.099
(3) BL 89.305 93.129 91.128
(4) BGA 90.12 91.589 90.789
(5) BLA 88.515 93.422 90.897
(6) BH 89.249 91.258 89.978
(7) BHW 92.815 88.985 90.844
(8) B_LEX 91.276 91.295 91.281
(9) BHW_LEX 90.166 91.699 90.861
(10) BERT-BGRU 87.965 94.669 91.276

Confusion matrix

Answer Positive Negative
Forecast
Positive) Kidney-YANG False resistance
Negative False Yin Kidney-YIN
Language:
English