Is Artificial Intelligence Intelligent Enough to Read Human Sentiment?

Abstract

Sentiment analysis is the mining of text which extracts information from the source material and helps a business to understand the social sentiment of their service while monitoring online conversations. Sentiment analysis can apply in any social domain because opinions are central to human activities and prominent influencer of our behaviors. Our beliefs, perceptions of reality, and choices are conditioned by how others evaluate the world. Due to this, when humans need to make a decision, they often seek out the opinions of others. According to GlobalWebIndex, 54% of social media users use media reviews, and 71% are more likely to purchase services based on social media referrals. Therefore, Sentiment analysis is "opinion mining" because it's all about digging into the context of social posts to understand the opinions they reveal. Misleading information has a more substantial effect on society than before. Not only this, but cognitive bias is also systematic errors in thinking, usually inherited by cultural and personal experiences, that can lead to distortions of perceptions when making decisions. Consequently, sentiment analysis can provide you with a broader perspective prior to making a final decision. This research aims to use artificial intelligence techniques for an in-depth analysis of human sentiments. The use cases and dataset used for this research are: 
 
 Stock Analysis using FinViz website
 Climate Change Analysis using Twitter 
 Movie review using IMDB dataset
 
 I used NLTK, Vader and Logistic regression machine-learning models and received approx 80% accuracy which shows us it is very promising that Machine Learning can help understand human sentiments and facilitates informed decision-making.