Of all the disruptive and groundbreaking features that Artificial Intelligence has presented us is Natural Language Processing (NLP). This subtype of AI extracts meaning from human language to make decisions based on the information. It focuses on the interactions between human language and computers. And, one of the best aspects of NLP is sentiment analysis. This is based on StanfordNLP, which can be used to identify the feeling, opinion, or belief of a statement, from very negative, to neutral, to very positive. Whether it is a judgment, texts, opinion, or emotional state, the primary goal is to identify subjective information in them.

This technology is commonly known as opinion mining or Emotional AI. It allows the study and development of devices and applications with an ability to recognize, interpret, process, and simulate human effects. Broadly, it comprises of two different types of analysis viz., Supervised, and Unsupervised sentiment analysis. In former, labeled sentences are used as training data to develop a model while the latter relies on the lexicon of words with associated sentiment values instead of labeled sentences. Because sentiment analysis can be automated, decisions can be made based on a sizeable relevant amount of data rather than pure intuition that may not always be right or prone to error.

There are three main techniques employed for sentiment analysis. These are:

Rule-based Method: It categorizes text based on unambiguous 'affect words' or words labeled by sentiment to determine the sentiment of a sentence. These sentiments are typically about emotions like happy, sad, rude, sarcasm, appreciation, and others. It is used for simple sentences or phrases like, e.g. I love rains.

Statistical Method: Here, an ML model is used to recognize the sentiment based on the words and their order using a sentiment-labeled training set. Under this, the model can detect the holder of a sentiment (the person with the opinion) and the target (the product or service) in a sentence. This is used for complex sentences like, e.g., Generally, I would not say I like rom-com.

Hybrid Method: This uses a combination of the above techniques, along with an advanced understanding of language to detect semantics that is expressed subtly, e.g. building relationships between implicit and explicit concepts.

Sentiment analysis is most helpful when deriving insights from large volumes of data at a shorter time, especially on customer data. This can help companies who want their brand to be perceived positively or have better responses and reactions in the market than its competitors. Another adopter for this technology is stock trading companies. Here, sentiment algorithms detect the companies that show a positive sentiment in news articles. This can mean a significant financial opportunity, as this may trigger people to buy more of the company's stock. Such insights help investors, traders, stock market participants in making decisions before the market has time to react.

Another great application is monitoring and analyzing sentiments of social media posts. A widespread use case is trying to predict elections based on the sentiment of tweets before the elections and results announcement day. This is possible as the most frequent kind of sentiment analysis performed is called polarity detection, i.e. understanding if a text about a given subject is positive, neutral, or negative. One can use sentiment analysis either through open-source tools, online APIs offered by Amazon Comprehend, Google Cloud; or SaaS products.

Sentiment analysis has gone through some significant highs and lows in the past few years. However, thanks to recent exemplary breakthroughs, we can make sense of human sentiments and emotional cues in real-time by training computer systems with supervised self-learning datasets. Thus helping in the evolution of this technique that has finally grown up into a core technology for the majority of today's NLP/NLU systems.