Emotional state recognition from text data using machine learning and deep learning algorithm

Abstract

AbstractEmotional state recognition is a process to identify user's feelings and emotions for various purposes. Emotional state examination from text comprises of extricating data about feelings, opinions, and even feelings passed on by scholars toward subjects of interest. Web‐based media is producing a tremendous measure of assessment rich information as remarks, notices, blog entries, and so forth. It is trying to comprehend the most recent patterns and rundowns the state or general feelings about items because of the enormous variety and size of web‐based media information, and this makes the need of computerized and ongoing conclusion extraction and mining. Sentiment analysis is difficult because of the existence of bad or abusive language with misspellings words. One of the major natural language processing research area is inclined toward understanding human emotions. Emotional state analysis acts like an amazing treasure and powerful tool, which renders its service to the field of deep learning. It can help service providers to fetch the requisite information to collect and identify the sentiments of the database. Principle issues that exist in the current procedures are: powerlessness to perform well in various areas, deficient exactness and execution in assessment examination dependent on lacking named information, inadequacy to manage complex sentences that require more than emotional words and basic examining. It is as yet hard for a greater part of instruments to decisively assess what genuinely is a negative, unbiased and a positive articulation. It is not advanced enough to successfully deal with sarcasm or context. So there is requirement to develop a machine learning algorithm to analyze the text data and give more better and accurate results.