-3

For a school project we have to find out if a tweet is related to fitness / physical activity or not. The whole dataset contains 50 million unlabeled tweets.

What are possible classifications for this undertaking?

Matthias Seidl
  • 29
  • 7
  • 1
    This seems far too broad a question; what techniques are you supposed to use - simple text matching, semantic analysis, Deep Learning, or other? – Ken Y-N Nov 08 '17 at 00:35
  • it doesn't matter which technique we use! it just has to be something more advanced then just writing a dictionary and checking if the words of the dictionary appear in the tweet or not. – Matthias Seidl Nov 08 '17 at 00:48

1 Answers1

0

One way to do this would be with some sort of artificial intelligence.

One of the most popular frameworks out there is TensorFlow by Google. I assume if you are doing this for school you have some experience with AIs anyway...

If you need some help getting started there is a very good YouTube tutorial series that is great for beginners which you can find here.

Good luck!

Trevor Clarke
  • 1,482
  • 1
  • 11
  • 20