Sentiment Analysis By Using Modified RNN And A Tree LSTM

In recent times, there has been an exponential growth in the number of complex documents, texts, and user-generated content, mainly through social media, blogs and review websites, etc., that require a deeper understanding of deep learning methods to be able to accurately classify texts. This research work approaches problems of fine-grained sentiment analysis. We explore how the sentiment of the user-generated text can be identified using Deep Learning. This paper contributes sentiment analysis, which aims to extract opinions from the text. E-commerce, social media, and other many online user-generated contents provide a powerful way for collecting people’s thoughts and their feelings towards users or customers. Powerful sentiment analysis helps many businesses with regard to sentiment, so they will understand users’ perceptions of the product. It’s also very helpful for other customers to get reviews of products and get to know about services they offer. Lots of websites provide users with reviews but it’s very difficult to make some decisions from them. Deep Learning sentiment helps to identify the pros and cons of a product. Nowadays Automated sentiment analysis systems are tools for decision-making for many parties. The deep learning algorithms are able to handle complex sequential data and determine non-linear relationships within data. Recurrent Neural Networks (RNN) are widely used in natural language processing (NLP) because they are very suitable to process variable-length text. Doing Fine-grained sentiment analysis is a challenging task. We use a modified Recurrent Neural Network (Tree-LSTM) for fine-grained analysis. In this work, we provide a brief overview of deep learning classification algorithms and also conclude that Tree-LSTM gives state-of-the-art accuracy for fine-grained sentiment analysis.