ML Study Round #9 – Kaggle

---Update October 2018: I made the format more structured----

Very cool, new location at CODE University this time ;)

We will focus on Kaggle challenges, because it's easy to see how well you're doing and nice to collaborate on.

For people wanting to do structured data:
https://www.kaggle.com/c/house-prices-advanced-regression-techniques
Feature engineering and figuring out what to do with data you don't know is probably more important than knowing the ins and outs of algorithms.

Unstructured data:
If enough ppl with image xp show up we can form groups around this one too:
https://www.kaggle.com/c/rsna-pneumonia-detection-challenge/data

If you don't like those, we can build groups around others – that's fine.
2-5 people per group is best usually

YOU HAVE TO TRY AND WORK WITH THE DATASET BEFORE COMING!

The main thing is to work together to make the best model, test assumptions and objectively see how well yo're doing.

----Check the page for setup and ML resources---

This meetup is about pragmatically getting better at machine learning by coding and experimenting. Don't fear, there is usually someone who can help you through the hard parts.

See you then,
Markus