Research Article
[Retracted] Development of Integrated Neural Network Model for Identification of Fake Reviews in E-Commerce Using Multidomain Datasets
Table 4
Comparing the results of an in-domain datasets with existing work.
| Paper id | Domain dataset | Features used | Method | Accuracy |
| Faranak Abri et al. [27] | Restaurant | Linguistic features from review content | MLP | 73% | Ren Y et al. [22] | Hotel | Review content and pretrained word embedding (bag of word) | CNN | 84% | Barushka et al. [33] | Hotel | Review content with pretrained word embedding (skip-gram) | DFNN | 83% | Garcia L. [24] | Amazon | Review content with TF-IDF | SVM | 63% | Hajek et al. [19] | Amazon | Review content with pretrained word embedding (skip-gram) | DFFNN CNN | 82% 81% | Barbado et al. [17] | Yelp | Review content with TF-IDF | AdaBoost | 82% | This study | Restaurant Hotel Yelp Amazon | n-grams of the review content with word-embedding matrix using embedding layer | CNN-LSTM | 77% 85% 86% 87% |
|
|