How Sentiment Analysis Can Analyze Mood

sentimental-analysis

COVID-19 And Sentiment Analysis 

2020 was a year that changed history with the spread of the COVID-19 pandemic. As the world shut down and people congregated online, social media channels lit up with collective moods and feelings about the pandemic. Historically, such data would have been subjective however with technological advances in text mining, it was not. Sentiment analysis has evolved in recent years with more uses in the business and public sector.

For the past 50 years, researchers have been working on technology that would enable computers to analyze the emotional tone of words. Such findings have had many real-world applications. For example, a large collection of tweets can be analyzed for collective mood something that would be impossible given human limitations. Now, music researchers can quantify how much a minor chord is sadder than a major chord. And businesses are tapping into websites like Yelp to understand customer behavior better. By assessing a vast number of reviews businesses can align brands with what customers feel and want. Sentiment analysis has medical applications as well with popular platforms like Facebook being able to recognize if users are depressed or suicidal.

When algorithms can gauge mood from what people write online it opens a new world of data and opportunity for society.

How Sentiment Analysis Works

How does sentiment analysis work? Advancements in Artificial Intelligence (AI) and Natural Language Processing (NLP) have mastered analyzing the context of language. However, AI has yet to understand the nuances of language. It was a significant win for science that artificial cognition can decipher the emotions behind words without understanding them.

In the beginning, sentiment analysis used word-counting for analysis. It added up the number of positive words and subtracted the negative words. Some content analytics platforms use weighted analysis to determine the mood of the content. For example, the word “atrocious” is given a higher weight than the word “bad”. Humans assign weights to words by creating

customized dictionaries within the software. In the sphere of sentiment analysis, these dictionaries are called lexicons.

However, human work is prone to error. Word-counting is fundamentally effective but ultimately unreliable. There are too many variables in language for word-counting to be precise. A more sophisticated approach to sentiment analysis uses Machine Learning (ML) algorithms to teach computers to recognize patterns and relationships between words. For example, a computer can recognize that “computer” and “mouse” are related and that it is a different type of “mouse” than the rodent variety.

Sentiment Analysis And Mental Health

Computer scientists are working on further advancements in sentiment analysis which has a foundation in psychology. As early as the 1960s, psychologists in leading universities were designing programs that were giving insights into individual’s psychological worlds. For example, computers were able to pick up on specific patterns for patients diagnosed with depression. Depressed patients were apt to use pronouns more often such as “I” and “me” as well as negative affect words or words related to death. These insights have empowered social media platforms to be proactive about user’s mental health. For example, Facebook has an algorithm that can identify suicidal users and if necessary, send the user helpline numbers.

With roots in psychology, sentiment analysis can give tremendous insight into the collective mood on social media. This has implications in the lives of the consumer and the individual as businesses become better at understanding consumer behavior. Also, individuals become engaged in social media platforms that can accurately decipher their mental health. It represents a shift in how much influence collective mood has on society.

Post This Article

Share on facebook
Facebook
Share on twitter
Twitter
Share on linkedin
LinkedIn

Related Articles