https://blog.datawow.io/what-if-your-computer-could-read-5dd7c03738af

Today we might type more than we talk, enormous amount of text data has been generated every single minute since digital era. Data is new oil (and even more valuable, for those who know how to blend it). However, a human brain is not sharp at reading large of amount of text. Our eyes are limited of two sides, which disallow a normal person to read more than 200–300 words per minute. According to “The Guinness World Record Book”, Maria Teresa Calderon can read more than 50,000 words per minute: 166 times faster than us (Maria could finish reading this blog, by end of our first two line of this paragraph). Would it be great if all of us can beat Maria in reading by a few line of Python code? The answer is yes! It would be great, with some help from free software fastText, and little understanding of Python scripts

The first task is basic Natural Language Processing (NLP) and apply some technique in “text classification”,which every line of text data can be categorised into assigned groups. For example, banking industry may offer countless of financial services which require written statement from customers. Text classification module (classifier) can read those statement, and categorise them into one of the bank services category. Anyway, machines need to be taught in order to understand human language.