Blog of Andrés Aravena
Dry-Lab:

Learning Machine Learning, session 2

15 October 2018

Keywords

The second session of the workshop focused mainly on collaboration tools that we will use, and we saw our first machine learning application.

Git and GitHub

Can Kandemir made a nice presentation, available on line. Faruk Üstünel pointed us to this Udemy online course. Their explanation is very clear.

For my side I showed some examples using RStudio git interface. In my experience it is easy to start with git using a graphical interface, but the real deal comes with the command line tools. It is a long subject, but we do not need to be experts.

There is plenty of material on line. Two pages that I personally like are “Introduction to Git” and “Getting started with Git”. The last one is a Reference card, that can be printed. The page is free but you have to register. Maybe that is not a good idea, so feel free to find other material, and share with the rest of us.

To begin, I’ve created a public repository for us. You can see it at https://github.com/dry-lab/learning-machine-learning. Go there, create an account, clone it, and start making mistakes.

Remember, this workshop is for everybody to learn, and the best way to learn is to teach. Believe me, I know a little bit about teaching and learning.

Kmeans

The first example we saw used gene expression data for E. coli, taken from Many Microbe Microarrays Database (M3D). We downloaded the last version, which has an approximate value of 2 million US dollars, although it is 10 years old.

We loaded this data in R and we did a quick clustering using the kmeans() function.

For the next session we have two challenges:

  1. Doing k-means in python
  2. Finding the “best value of k

See you there!

Originally published at https://anaraven.bitbucket.io/blog/2018/dry-lab/workshop-session-2.html