Senti_ALSTM: Sentiment Analysis of Movie Reviews Using Attention-Based-LSTM

Association with the customers is one of the principal elements for businesses. It is essential for these firms to recognize exactly what clients’ opinions about the latest as well as established products and services. One such business is film industry. Movies are a dominant source of entertainment. Yield of a movie relies on a great extent to the reviews on social media, blogs, forums, IMDB, and movie viewing platforms. In this paper, to achieve revolutionary performance in sentiment analysis, attention-based long short-term memory, ‘Senti_ALSTM,’ model is proposed which improves the neural networks by assigning memory to it and further reducing the vanishing gradient problem of recurrent neural network. Senti_ALSTM experiment with attention in long short-term memory model which takes into account the information reflected by the input sequence, thus remembering the context, and establishes the relation between input and output sequence. Further, the success of the benchmark data confirms the efficiency of the proposed model over the other models.