Review Rating Prediction Using Attention Bi-directional Gated Recurrent Unit with Domain Based Word Embedding

Review sites allow the customer to express their product reviews and provide ratings for all products. Because of the massive development of online reviews, customers are struggling to find better-quality products. To overcome these issues, different models with respect to the rating of users are presented. However, these approaches have incorrect rating predictions since it is complex to find the preferences of users. Enhancing the performance of ratings obtained from the recommendation is important. This work presents an ABi-GRU (attention bi-directional gated recurrent unit) with the DBWE (Domain based word embedding) for RRP (review rating prediction). This work used DBWE and it is not based on the information of the customer and the ABi-GRU is used for capturing the hidden context information. The experimental analysis is demonstrated on the two benchmark review datasets. The findings proved that the performance of the proposed DBWE-ABi-GRU achieved better RRP outcomes.