Natural Language Processing and Social Media

In this article, we will talk about Natural Language Processing (NLP).

This is what you can learn here:

• What is Natural Language Processing?
• What are the most popular programming languages?
• Where we use them and what is their contribution?
• What are the most popular applications and what is their usage in different fields?
• The NLP example on Twitter
• Is NLP ever going to replace human support?

🚀 Download our free eBook: Create a chatbot that delivers!

What is natural language processing 

Natural language processing or NLP for short is often described as automatic manipulation of text or speech specialized software. It is an AI branch that assists computers in understanding and interpreting human language.

While NLP isn’t a new science, the technology is still rapidly progressing because of the increased interest in human-to-machine communications. The computer’s natural language is computer code or machine language. It is incomprehensible to most people because communication occurs not with words but through millions of zeros and ones that generate logical actions to the machine. That kind of communication is comprehensible to a relatively small amount of people.

Why is natural language processing important

NLP is used in many tools that we utilize on a daily basis. It is a driving force behind applications like google translate, Microsoft word, Grammarly and many more. It’s even responsible for personal assistant apps like Siri, Cortana, Alexa, and others.

The Technical Side of Natural Language Processing

The most popular NLP programming languages are:

• Programming language R
• Python
• Java

1. Programming language R

Programming language R is used for statistical learning. It understands and explores your data using statistical methods and graphs. R plays an important role in investigating big data, supporting the researcher, and also it is useful for intense learning analytics. It has an enormous number of natural language processing algorithms.

Picture 1. Why learn R?

2. Programming language Python

Another one is Python. It’s a high-level object-oriented scientific programming language. Python is easy to learn syntax readability and reduce the cost of maintenance. It contains a lot of packages using this you can do code reusability.

Also, it can extract information from unstructured text, either to guess the topics and identity named entity. Using Python, parsing and semantic you can analyze language structure.

3. Programming language Java

The most popular programming language for Android Smartphone is Java. Java helps you to organize text using full-text search, clustering, tagging, and information extraction. It is a platform independent language because of this feature the processing of information becomes easy.

There are some other applications of this new technique:

Grammarly is an enhancement for grammar checking.

Twinword Writer is a writing platform. It’s also used to translate from one human language to another.

Computer translations cut down the time needed for translating documents.

More demanding applications of NLP are chatbots and humanoid robots.  Chatbots or intelligent agents answer all types of questions and inquiries from their users.

Humanoid robots, such as gynoids, use NLP as their base of understanding and responding to different stimulants.

Natural Language Processing in Social Media

The development of social media has revolutionized the amount and types of information available today to NLP researchers. Data available from social media such as Twitter, Facebook, YouTube, blogs, and discussion forums make it possible to find relations between demographic information, language use, and social interaction.

Using statistical and ML techniques, researchers can learn to identify demographic information, language, track trending topics, and predict disease spreading. For instance, with Google Flu Trends it recognizes deception in fake reviews from symptoms mentioned in a tweet or food-related illnesses.

Social media has changed the ways we can use big data. Product reviews can help to predict pricing trends and plan future advertising campaigns. Political forums help to predict success in elections.

Equally important, social networks can be used to determine indicators of influence among different groups. Similarly, medical forums can contribute to the discovery of questions about patients suffering from a particular medical condition.

Related: Importance of Social Media Monitoring

Walt Disney Example

Here’s the tour of how a simple NLP project is structured and used to gauge public opinion of Disney in real-time, at a minimal cost.

Picture 2. Alice in Wonderland

This research was based on around 100,000 to 120,000 tweets related to Disney, collected each day, in over 30 different languages. The tweets were run through a Python library called TextBlob, which analyzes each data point of the text, in this case, a single tweet, and calculates the Polarity. 

This is how and why certain steps are done along the way.

The first source of bias was discovered while exploring the data. While searching for many permutations of keyword Disney (i.e. “Disney”, “Disney’s”, “#disney”, etc.), it did not collect all possible tweets related to the topic, but only tweets that mention one of the words in the search list. 

As it turns out, The Disney Company is really good about taking intellectual property and spreading it across many of its businesses. Not only that, but it seems that when somebody loves a Disney movie, they will mention it all in a single 280 character tweet.

The final example:
– “Disney Is Not Only Disney Parks! #Disneysmmc @DisneyCruise@DisneyAulani @DVCNews @Disneymoms”

In this single tweet, there are the following keywords: ‘Parks,’ ‘Cruise,’ ‘Aulani’ (Disney resort in Hawaii), ‘smmc’ (a special event at Disney — Social Media Moms Celebration), and ‘DVC’ (Disney’s timeshare program — Disney Vacation Club).

As with the previous example, this is interpreted as a lot of potential signal in a tweet, but not enough distinct signal for the model to be able to place it in a single category with textually similar tweets of the same topic.

Focus on Sentiment Analysis

What is more important, much of the work in NLP has focused on sentiment analysis.

The sentimental analysis determines opinions, committed or uncommitted beliefs, emotions, positive or negative orientation.

In the same way, it determines the connotation of textual language or speaking language. All mentioned is based on lexical and syntactic information.

Positive and negative orientation signal by words sentimental attitudes. For example, words with negative connotation are sad, worried, difficult, and weak. In contrast, comfortable, important, successful, and interesting convey a positive sentiment.

Online sentiment dictionaries, such as Whissell’s Dictionary of Affect, can assess positive and negative sentiment in text. However, there are new practices for identifying emotions based on six basic ones.

There has also been researching using features that can recognize classic emotions to identify speaker medical conditions.

For instance, it can recognize autism and Parkinson’s disease. In addition to medical conditions, it can recognize speaker characteristics such as age, gender, likeability, pathology, and personality. Moreover, it can recognize the speaker’s conditions like cognitive load, drunkenness, sleepiness, interest, and trust.

What Is Next for NLP?

Recent advances in Machine Learning (ML) together with deep learning have enabled computers to do quite a lot of useful things with NP.

Furthermore, it has enabled us to write programs to perform things like language translation, semantic understanding, and recognizing emotions. Although there is a disadvantage where computers don’t yet have the same intuitive understanding of natural language that humans do.

Because of that, it’s impossible for them to “read between the lines”. That’s why it’s justifiable to doubt that they won’t be able to do a better job than humans.

But, if we reconsider that humans also often are not able to “read between the lines”, can handle less information and are slower –  NLP is a great opportunity to take over some humans responsibilities.