Tweet Generator was an extremely fun project to work on. This project implements a basic form of Natural Language Processing utilizing Markov Chains. Markov Chains are structured similarly to a finite state machine model.
The way the Tweet Generator works is divided into steps:
-
Initally, the program will take in a given text file and build a Markov chain using a
Dictogram Structurewhich is simply a Python's dictonary class extended to have additional functionality. -
The chain is built by specifiying an order. The order indicates the amount of words in a given state.
For Instance take the sentence
one fish two fish red fish blue fish
A second order chain would dictate that each state would be two words one fish fish two two fish etc..
each state contains all words that follow that exact sequence of characters. The higher the order the more consecutive words appear in a single state. Too high of an order will only generate sentences that already exisit in the sample text.
-
As the chain is built start and stop states are generated. Start states consists of the begining of sentencres and stop states consists of ending punctuation like
. ? !. -
Now the fun begins! This is the point where the chain is fully built and the sentence starts to generate. Markov chains use a form of stoicastic sampling which just simply means words are chosen based on probability.
- Audaris 'Audi' Blades
Simply click here to go to the tweet generator's website. The server may take a minute or two to load up as it switches from a dormant to active state.
- Audaris 'Audi' Blades
- Python - Programming Language
- Flask - Lightweight web application framework
- Jinja - Template engine for python
- Bootstrap - Front end framework
- Heroku - Application deployment site
- Make School's CS 1.2 course provided the fondation for this assignment.