An improved gated recurrent unit based on auto encoder for sentiment analysis

Sentiment analysis is a particularly common task for determining user thoughts and has been widely used in Natural Language Processing (NLP) applications. Gated Recurrent Unit (GRU) was already effectively integrated into the NLP process with comparatively excellent results. GRU networks outperform traditional recurrent neural networks in sequential learning tasks and solve gradient vanishing and explosion limitations of RNNs. This paper introduces a new method called Normalize Auto-Encoded GRU (NAE-GRU) to address data dimensionality reduction using an Auto-Encoder and to improve performance through batch normalization. Empirically, we demonstrate that with slight adjustments to hyperparameters and optimization of statistic vectors, the proposed model achieves excellent results in sentiment classification on benchmark datasets. The developed NAE-GRU approach outperformed other various traditional approaches in terms of accuracy and convergence rate. The experimental results showed that the developed NAE-GRU approach accomplished better sentiment analysis accuracy of 91.32%, 82.27%, 87.43%, and 84.49% on IMDB, SSTb, Amazon review, and Yelp review datasets respectively. Furthermore, experimental results have shown that the developed approach is proficient in reducing the loss function and capturing long-term relationships with an effective design that achieved excellent results compared to state-of-the-art methods.