Skip to content

Latest commit

 

History

History
41 lines (24 loc) · 2.5 KB

README.md

File metadata and controls

41 lines (24 loc) · 2.5 KB
Raw

Udacity_NLP

This repository was made to store materials for Udacity Natural Language Processing NanoDegree Program, which I'm currently enrolled in.

Disclaimer

Every notebook and iamge/diagram included in this repository is from the Udacity Natural Language Processing NanoDegree Program.

Natural Language Processing?

Arguably, Natural Language Processing is one of the most hyped and well-recognized branch of Artificial Intelligence discipline. It aims to equip machines to be able to understand the texts and contexts we use, which is a whole difference level of subject than basic machine learning discpline.

This is because the meaning of contexts can be drastically changed only by a single word or by where a specific word positions in a sentence. It can be even more sophisticated by how puntuation marks can shift the meaning with that comprise same words.

Because of this, advanced technologies, such as deep neural networks are often used to make use of good understanding of human languages. Rather than relying on current timeline data, Deep neural networks enable computers to employ wider range of timeline data (e.g., past inputs) to make a good predictions.

Natural Language Processing Applications

There are a number of applications of this cutting-edge technology, and the following are only a part of them.

  • Machine translation (Google Translate)
  • Voice recognition (Alexa, Siri, Google Assistant)
  • and more ...

Basic steps in Natural Langauge Processing

Regardless of applications, preprocessing step is needed for the purpose.

  • Normalize (lowercase the capital letters)
  • Tokenize (split words into separate tokens)
  • Remove stopwords (punctuation marks are removed for core text processing)
  • Stemming/Lemmatization (verbs are converted to root forms): This step is often omitted for applications such as Machine Translation

My intentions to move forward

Why did I set my mind to move forward with this discipline?

I was motivated with a podcast episode of NPR's Platnet Money, BOTUS. BOTUS means Bot of the United States, just like POTUS. That implies what this bot does.

Namely, it analyzes President Trump's tweets and makes trading predictions only relying on them. This is clearly an application of Natural Language Processing, and I was intrigued by that and I hope to see myself implement something like that in the near future.