Word association algorithms are an essential part to any machine learning algorithm dealing with writing. It serves as a dictionary that is understandable by the computer. These word associations, however, do not come from a predetermined list, but rather from regular sentences. Take this sentence for example: 

 The yellow bus pulled into the bus stop to pick up people. 

 Analyzing the word “bus,” we can infer that the bus is yellow and that its purpose is to pick up people. Obviously to us, we can understand a bus’s purpose, but to a no-nothing computer, this information is very important. 

 Below is a version of my software that is web compatible. The text that we are analyzing is from a Wikipedia article about Australian Rules Football (Link: https://en.wikipedia.org/wiki/Australian_rules_football). 

 Just a fair warning, because this software does many calculations per second, it does cause computers to heat up considerably (as I learned from my poor computer). I would recommend that you only run it for a little bit and stop it if you have to change tabs.

Iterations

(This is the number of times the computer has evaluated a word):

Loss

(Loss is the measurement of how off the computer is from the true answer. The lower the number, the better):

Word

(Shows the current word being evaluated):

Generated Vector

(This is the group of numbers that represent the word):

Word Cloud

(This is a visual representation of how the first 2 numbers group data. I cannot show all the numbers at once because that would require more than 3 dimensions, which is impossible. I swear I’m not being lazy):