Sentiment classification is a popular technique in natural language processing (NLP). Sentiment classification is referred to as the automated process to identify the opinions of the users from their text and label them as positive, negative, or neutral based on the user-s emotions in their text. Contradictional texts or negation terms also contain valuable information to determine the polarity of the user-s view, and the sentiment analysis ignores such text.
Traditional negation handling techniques are failed to address the user-s contradictional texts accurately. Contextual representation of user-s text helps the sentiment classification to handle the negations without canceling it properly. Contextual representation focuses on a complete analysis of user-s text with its neighborhood and genre information. Contextual information plays a significant role in decision-making for polarity computation. Negation handling for sentiment classification using contextual representation effectively determines the polarity of user-s emotions in their text and improves classification accuracy.
• In natural language processing, sentiment classification has gained far-reaching acceptance in recent years.
• Sentiment classification recognizes and extracts instinctive information of opinion polarity from the text.
• Negation handling is the crucial sub-task in sentiment analysis.
• The inability to precisely resolve the negation effect is one of the core reasons for error in sentiment analysis classification.
• Contextual representation effectively handles the negation, as the impact hybrid approach with the lexicon model is utilized for sentiment classification.
• Deep learning models are also employed for sentiment analysis to learn complex syntactic features of negation automatically.
• Recently, deep intelligent contextual embedding has been presented based on word representation by encoding representations from the transformer to enhance sentiment classification.