Many machine learning beginners are confused about how to approach machine learning. Machine learning is huge. As a beginner if you do not follow the right approach, those integration symbols and probability formulas will scare you to death and you will soon lose interest in the subject.In this article I will explain how to learn machine learning the easy way and common pit falls that a beginner should avoid while starting off with this course.
Many will advice you with what are the good books as a beginner you need to study to get yourself up to speed. My advice is Do not start with a book. If you are not from mathematical background, it will scare you to death. Start with a online course in Udacity, Course era and Course on you tube by Yasser Abu-Mostafa. My advice would be to complete all these three courses with the same following order
- Course era (Andrew Ng)
- Course on you tube by Yasser Abu-Mostaf
The first course in Udacity will directly pull you into the practical field with real world example. It will also explain you the required theory. This is an excellent course to get started with. The course by Andrew Ng will go into little more depth into the theory and finally Course by Yasser Abu-Mostafa will teach you a good amount of theory. Before starting the course on Udacity, get yourself familiar with little bit of python. It's a very easy language to learn and you will only need to know the basics of python to complete this course. The course uses scikit-learn which is a python based software package for machine learning algorithms and its very powerful. I use scikit-learn myself for all my machine learning work. However some of my peers also use R
Once you are done with these courses, please don't open a book yet. Open an account in Kaggle. Kaggle is a platform for machine learning competition. Try out some easy old competition. You will find solution with complete code in youtube. This will really get you going and will boost your confidence to a very high level. After that, compete in some live Kaggle competition. Check out the benchmark codes. Try to improve the bench mark. Read the comments for that competition. Create a group and participate as a group.
After all these, chose a topic. For example - SVM, Deep Learning, Ensembling etc. and now pick a good book for that particular topic and master the theory behind that algorithm.
From this point onwards, you will no longer need an advice. Best of luck.
Sharing is Caring!
RECOMMENDED POSTS FOR YOU