An Effective Method for Text Sentiment Analysis using Emotions

Abstract :

Sentiment analysis refers to the task of natural language processing to determine whether a piece of text contains some subjective information and what subjective information it expresses, i.e., whether the attitude behind this text is positive, negative or neutral. Understanding the opinions behind user-generated content automatically is of great help for commercial and political use, among others. The task can be conducted on different levels, classifying the polarity of words, sentences or entire documents. Deep learning for NLP or Deep NLP model is used to classify the sentiments of the feedback. In sentiment classification model, certain limitations existed in the current method, which motivated us to plan for developing this research proposal. To compare the fundamental performance of deep-learning models, words and characters were randomly initialized and trained together with network parameters in current system, this arises the problems like sarcasm, tone, comparison, polarity, etc. However, certain Word vectors, Dynamic masking pattern, Aspect-based Sentiment Analysis approaches and Full sentence predictions are beneficial for avoiding current method problems and help to enhance accuracy and performance of the text sentiment analysis.