Biggest Mistake I Made When Starting Machine Learning, And How To Avoid It

 When I first got started in machine learning I implemented algorithms by hand. It was really slow going.

I was a terrible programmer at the time. I was trying to figure out the algorithms from books, how to use them on problems and how to write code – all at the same time. This was the biggest mistake I made when getting started. It made everything 3-times harder and killed my motivation.

A friend of mine suggested I look at Weka. I resisted. I was a student and a C-programmer and I didn’t want to get into Java.

Later, I was looking into decision tree algorithms and I learned that Weka had an implementation of C4.5 (a really powerful method).

I downloaded Weka and discovered that in addition to the Java API, Weka had a fully interactive graphical interface for loading data, running algorithms and reviewing results. Basically, all of the things I was trying to figure out how to do and implement myself, but in a GUI.

Kick-start your project with my new book Machine Learning Mastery With Weka, including step-by-step tutorials and clear screenshots for all examples.

Weka Explorer Interface with the Iris dataset loaded

I was hooked. I started using it for class work and for my own experiments. Later in grad school, I started my own research by writing 3rd-party plugins for Weka (LVQ algorithm and others).

I now recommend it to programmers just getting started because it’s so quick to get meaningful results on a dataset. It also instills best practices like repeatable experiments and statistical methods for comparing results.

If you haven’t already, take a look at Weka.

I have a short tutorial in which you discover how to run a classifier in 5 minutes.

Good luck machine learning!