I made a Donald Trump simulator that takes what the man himself

File: Screenshot_2017-07-07-23-12-48.png (98KB, 1079x391px) Image search: [Google]

98KB, 1079x391px

I made a Donald Trump simulator that takes what the man himself said and uses markov chains and grammatical data (not that the man tends to use much of it) to make word connections of his tweets, then proceeds to generate its own sentences and posts it to Twitter every 5 minutes or so.

The script takes into account every tweet of his and also all previous tweets itself posted, and that way it will never post a exact duplicate. furthermore, it makes sure to "learn" by distance itself from previous tweets (essentially merging the 2 separate chains with a less than 1 multiplier, I use 0.5. So every time a word connection is used it is half as likely to be used the next time around)

So every tweet is unique, mostly follows grammatical rules and reflect what trump could have said.

Head over to @DonaldTrumpSim on Twitter if you are interested to see what all of that logic above boils down to.