Table 4.
Topic modeling methods | Promotion topic | Review topic | ||||
|
Precision | Recall | Precision | Recall | ||
MALLETa document classification, mean (SD) | ||||||
|
NaiveBayes | 81 (2) | 80 (3) | 64 (2) | 87 (3) | |
|
C45 | 66 (7) | 68 (11) | 56 (6) | 75 (9) | |
|
Decision tree | 83 (3) | 51 (3) | 71 (2) | 61 (4) | |
MALLET topic modeling, n (%) | ||||||
|
Unsupervised topic modeling | 814 (48) | 814 (81.40) | 939 (53.69) | 939 (93.90) | |
Our model, mean (SD) | 88 (1) | 85 (2) | 82 (1) | 80 (1) | ||
Baseline, mean (SD) | 84 (1) | 84 (3) | 76 (3) | 74 (2) |
aMALLET: Machine Learning for Language Toolkit.