|
Sentiment analysis using attention-based convolution autoencoder (SAABCA) |
|---|---|
| รหัสดีโอไอ | |
| Creator | Neha Sharma |
| Title | Sentiment analysis using attention-based convolution autoencoder (SAABCA) |
| Contributor | S Veenadhari |
| Publisher | Rangsit University |
| Publication Year | 2565 |
| Journal Title | Journal of Current Science and Technology |
| Journal Vol. | 12 |
| Journal No. | 2 |
| Page no. | 224-242 |
| Keyword | Sentiment Analysis, Deep Learning, Autoencoder, Attention based Convolution Autoencoder |
| URL Website | https://jcst.rsu.ac.th/ |
| ISSN | 2630-0656 |
| Abstract | Sentiment analysis (SA) has been a commonly studied issue in the disciplines of NLP, data mining and data analysis. Deep neural network (DNN) algorithms have lately been used to do SA with significant improvements. However, these algorithms can handle sequencing of any lengths, employing it in the extracting features of a Deep Neural Network increases the dimensionality of the feature space. In this paper, a sentimental analysis using attention-based convolution autoencoder (SAABCA) model is proposed to tackle such challenges. Moreover, the Attention mechanism (AM) is used on the outcomes of layers. SAABCA employs convolutional and pooling methods to decrease feature dimensions and recover position-invariant feature points. The efficacy of model is measured by its ability to identify sentiment orientation, that is the most popular and essential job in SA. As contrasted to seven previously recommended DNNs for sentiment analysis, additional state-of-the-art effectiveness is shown on several review classifications and tweet polarity classifications. The findings indicate that the suggested technique accuracy is 94% for Kindle Dataset, 97% for Movie Dataset and 98% for Airline Twitter Dataset. |
