Research Article
Uncovering Cybercrimes in Social Media through Natural Language Processing
Table 4
Sentiment analysis results for Scenario 1 using SLP algorithm versus Vader model.
| Sentiment analysis algorithm | Cluster | Polarity | Subjectivity (mean) | Number of tweets | Number of accounts | Negative (%) | Neutral (%) | Positive (%) |
| SLP algorithm (TextBlob library) | 1 | 29.8 | 42.1 | 27.9 | 0.29 | 368 | 302 | 2 | 13.6 | 59.1 | 27.2 | 0.20 | 88 | 73 | 3 | 17.3 | 57.1 | 25.5 | 0.22 | 196 | 167 | 4 | 35.5 | 22.2 | 42.1 | 0.43 | 453 | 362 |
| Vader model (VaderSentiment library) | 1 | 49.4 | 23.3 | 27.1 | — | 368 | 302 | 2 | 18.1 | 59.0 | 22.7 | — | 88 | 73 | 3 | 40.8 | 36.2 | 22.9 | — | 196 | 167 | 4 | 54.8 | 8.1 | 36.6 | — | 453 | 362 |
|
|