Hacker News Reader

So what generalizations can we make from these plots? Instance and rule based methods learn better? Random forest seems to be especially reliable.

I don't think any useful generalizations can be made without information about the CPU and memory usage of the algorithms.