How to Get Started with Machine Learning
Machine Learning can be intimidating and learning it online could be overwhelming because there are tons of material out there, most people don’t know where or how to start.
In this article, I present a plan to self-learn Machine Learning that worked for many people I personally know including myself. It’s also important to note that there is no one size fits all learning path as each person is unique and have different goals that evolve with time. Treat this plan as a guideline or someone else’s experience that you could tweak to suit you.
Learn The Basics
If you don’t have the basic mathematics and statistics behind Machine Learning then it’s a good idea to build one now. Even if you have a mathematical background it’s always good to get a refresher. The amount of effort to put in this step is up to you.
One of the best ways to start learning the mathematics behind Machine Learning is the old fashion way of reading a book. In my opinion, one of the best books out there is An Introduction to Statistical Learning by Hastie et al. The good news is that you can download it for free thanks to the University of Southern California, School Of Business. Here’s the link to the book. Of course, you can buy it from Amazon as well.
Another very good book by the same authors is The Elements of Statistical Learning. You can download this book for free through this link, thanks to Stanford University. Also, you can buy it from Amazon.
If you’re not the type of person to sit and read a book then watching a video is a good option for you. Dr. Abu-Mostafa from the California Institute of Technology (CalTech) has a great free online course that attracted more than 7 million viewers. His course Learning from Data is 18 lectures divided into theory, technique, and analysis all lectures are on YouTube, here’s the link to the course.
When it comes to online Machine Learning basic courses there is the famous Stanford offered course by Andrew Ng on Coursera. This is the most common beginner course out there. It’s an easy to follow course with some quizzes, upon completion you will get a Coursera certificate that you can share on your Linkedin profile.
Programming Languages Choice
While you can use any programming language to write Machine Learning codes, Python and R are the most common ones. There are so many libraries and packages in Python and R specially built for Machine Learning. Also, you will easily be able to find community support through sites like StackOverflow. It’s recommended to stick with these two languages for Machine Learning.
But what if you don’t know Python or R or maybe you need a refresher. The good news is that there are several online courses that teach both Python and R with an emphasis on Machine Learning. Instead of spending hours learning the syntax of Python and R, you can take a crash course with a focus on Machine Learning.
My three favorite courses are Python A-Z: Python For Data Science with Real Exercises and Python for Data Science and Machine Learning Bootcamp which are based on Python language. R Programming A-Z: R For Data Science with Real Exercises, which is based on the R language. These courses will introduce you to the programming language (Python or R) and walk you through the main packages needed for Machine Learning along with how to use them.
Hands-on Machine Learning Courses
By now you learned the mathematics behind Machine Learning and built a strong foundation in Python or R or maybe both. Let’s put all this learning into action, it’s time to take an online course where you hands-on write codes to solve Machine Learning problems.
Again, there are tons of online courses out there so do your research and choose one the suits you. One of the courses I truly enjoyed is Machine Learning A-Z: Hands-On Python & R In Data Science. The reason I enjoyed this course is that each section is broken into 3 parts, first, the instructor explains the theory and presents a problem to solve, second, the instructor solves this Machine Learning problem using Python code, and finally again solves the same Machine Learning problem using R code. So it doesn’t matter if you’re programming with Python or R because this course teaches you to solve Machine Learning problems using both. Another hands-on course I enjoyed is Machine Learning, Data Science, and Deep Learning with Python, a lot of theory and Machine Learning practice coding exercises.
Do some Machine Learning Projects
At this step, you have a strong foundation in Machine Learning and Python or R. The best way to take your skills to the next level is by doing some real projects. Go to Kaggle’s Competitions page, you can start by looking at completed competitions and follow the attached notebooks. Whether you feel confident in your skills or not, I strongly recommend joining an active competition. In my opinion, joining any active competition is the best way to learn.
If you’re the kind of person who wants to take a class and run through some real-world Machine Learning problems then I recommend Machine Learning Practical Workout | 8 Real-World Projects. It is a hands-on class I found really useful to put your Machine Learning skills to work.
Earn a Certificate
It is now time to earn a certificate and put it on your LinkedIn profile. You want your colleagues, recruiters, and potential employers to know about you’re Machine Learning skills.
Again, this is my personal opinion so do your research. I believe the best Machine Learning Certificate out there is the AWS Machine Learning Specialty. You have to know basic AWS cloud to start studying for this certificate, it’s a tough certificate to earn and needs a lot of preparation. I wrote an article a while ago on 5 main concepts you need to understand to pass the AWS Machine Learning certification exam.
Another valuable certificate you could get is Udacity’s nano-degree Machine Learning program. Udacity partnered with AWS and Azure to offer Machine Learning programs geared towards these cloud providers. Check both the AWS-Udacity Machine Learning program and the Azure-Udacity Machine Learning program. These are months-long programs that teaches you Machine Learning, and AWS or Azure cloud fundamentals.
Finally, I summarize the learning path described earlier in the below table.
This worked for me and some colleagues, I’m putting it out there as a reference guide for you. It’s very important for you to design your own learning plan based on your goals, needs, and available time. What worked for someone may or may not work for another person, we are all unique but its good to learn from one another and see what worked and what didn’t. Good luck with your Machine LEarning journey, it’s a tough yet very rewarding one.
