In this project, I built two tweet generators and two tweet classifiers, and used the better-performed classifier to evaluate the performance of the generators. For the generators, I used about 24k tweets of Trump to train the tweet generators using Multinomial Naive Bayes and LSTM models.
Then I added another 24k tweets from Obama, Hillary, and Kim Kardashian and trained the tweet classifiers using Markov Chain and LSTM models.
Finally, I compared the classifiers and used the better-performed one to evaluate the performance of the tweet generators and output some Trumpy tweets.
-
For the python code, check this notebook. Data can be found under the /data folder.
-
For the details of this project, check this medium post.
-
For more details of this project, check this formal report.