Twitter is the Sentiment Analysis Hub for Research; it Predicts the Stock Market

Twitter certainly has a long future ahead for them in the tech world. It has performed quite well as a measure of public sentiment. Researchers found Twitter to be as useful as the Gallup Poll. In fact, many experts will go as far as to say that Twitter can predict the Stock Market. All you need are the right tools to discover these clues.

There are particular programs used by professionals, such as hedge funds that will aggregate certain keywords that come from the Twitter feed. The keywords will be further subjected to Boolean algorithms that can also remove excessive repeat words.

Once all the words are run through software filters, they can be categorized into attitudes like, Buy, Hold, Sell. It’s all based on what impressions are being communicated through Twitter.

The Twitter messages are all automatically processed. No human is needed until everything is categorized. But, this process doesn’t take long anyway. A large amount of information passes through Twitter. Catching all the key words is impossible by humans, or at least impractical.

This is assuming that the formula is even running correctly at all. From time to time, it will need to be tested and corrected. People’s language changes frequently in conversation. Sometimes new words quickly become popular which can be helpful for a business to find out quickly. Companies need to research and compile information as accurately as they can to gauge the meaning of it.

What are people’s attitudes towards: Movies, Products, Politics? It’s not easy to determine. Sometimes, people may be a little sarcastic about something and that will result in a bad data point, which will be in contrast to the real attitude. Even though, sarcasm doesn’t occur as much as truthful honesty. People tend to want to tell the truth.

Other words use to describe this process are: opinion extraction, opinion mining, sentiment mining, subjective analysis.

The human psychology experiences a wide range of emotion on a daily basis. At this time, the researchers are unable to make use of our emotions for credible dats points. The technology isn’t there to effectively react properly enough to our emotions. That also goes along with mood as well.

The computers also need to factor out our interpersonal stances. Such as, being cold to an enemy or warm to a friend. Computers can’t tell if you are supportive of another human on a personal level. Besides, the language is simply hard to communicate between humans sometimes. For example, many times a female or male may not even know if he was being flirted with.

Computers can’t detect personality traits effectively enough either. Your behavior tendencies may be hard to communicate anyway. Not many people can tell if you are hostile, anxious, nervous or jealous unless you go out of your way to make it clear.

The human trait that is most popular to measure these days is attitude. A good definition for attitude is: enduring, affective lay colored beliefs, dispositions towards objects or persons. Because this human behavior is more longer lasting, the scientists have put a lot of focus on it and how to create the perfect algorithm to find the right combination of words.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s