Machine Learning Education: 3 Paths to Get Started
Machine learning is the predictive heart of big data analytics, and one of the key skills that separates data scientists from mere analysts. But getting started with machine learning can be a challenge. Here are a few ways beginners can get off the ground with their machine learning adventure.
Machine learning is a vast field with many different specialties, so it’s quite easy for a beginner to get overwhelmed. For instance, one specialty called deep learning powers many of today’s artificial intelligence breakthroughs. But without a background in basic machine learning approaches, a prospective data scientist would have zero chance of mastering this powerful but complex technology.
Here are three general paths that neophytes can get started with machine learning:
Take a Course Online
Udemy offers a beginner’s course in machine learning called “Machine Learning A-Z: Hands-On Python and R in Data Science.” The course costs $49 and is open to people who have a grasp on high school-level math.
Udacity has a course titled “Intro to Machine Learning” that covers the basics, such as extracting data, model selection, and performance evaluation. The self-paced course is free of charge and takes about 10 weeks.
While it may not be suitable for beginners, Coursera‘s machine learning class taught by renowned data scientists Andrew Ng is regarded as one of the top machine learning classes around. The course takes about 11 weeks and goes into considerable depth into machine learning topics. Students can access course material for free, but will need to pay to get a certificate upon completion.
If you know R and statistics, but are completely new to machine learning, you might want to consider DataCamp’s “Introduction to Machine Learning.” Students can start the course — which covers basic ML algorithms like classification, regression, and clustering – for free, but getting a certificate will require payment.
Read a Book
If you want an in-depth descriptions into applies statistics, you could read “An Introduction to Statistical Learning with Applications in R,” which is available as a free PDF download. The 440-page book from Springer is suitable for those with mathematical backgrounds as well as beginners from social sciences, although it assumes you have taken a class in basic statistics.
No treatment on this topic would be complete without a book from the “Dummies” series, and we don’t disappoint. “Machine Learning for Dummies,” which is available at Amazon for about $18, tackles the topic from both R and Python points of view, although readers don’t have to have any computer science background to get something out of it.
The bestseller in this category may be “Machine Learning for Absolute Beginners: A Plain Introduction to Supervised and Unsupervised Learning Algorithms” ($10 at Amazon). Author Oliver Theobald covers the basic algorithms at play in this short, 132-page book, as well as some related facts, such as salaries that data scientists can expect.
Christopher Bishop’s 2006 “Pattern Recognition and Machine Learning,” also from Springer, is a highly regarded tome on the topic. Bishop’s 758-page book goes in-depth into the theory behind the technology, including Bayesian probabilities and Gaussian distributions. He also covers some graphical models. It’s not necessarily for the beginners though, as he assumes the reader has a firm grasp of calculus, linear algebra, and probability theory.
Watch a Video
You can access the video portions of Udacity’s free online “Introduction to Machine Learning” course (mentioned above) via YouTube, which they are the top-viewed machine learning videos. Your host in the videos is Sebastian Thrun, who’s the founder of Udacity and also a Google Fellow and Stanford professor.
Another source of machine learning videos is Kevin Markham, the founder of Data School and a former data science instructor at General Assembly and Coursera. Markham’s Data School videos focus on learning to use the scikit-learn library for Python. No data science or statistical experience is expected.
Josh Gordon’s “Machine Learning Recipes” video series offers a good introduction to machine learning. The Google Developer videos focus on using scikit-learn and Google’s own TensorFlow libraries to build basic machine learning models.
Andrew Ng’s Stanford lectures on machine learning remain some of the most-viewed YouTube videos on the topic of machine learning; his first CS-229 lecture from way back in 2008 has more than 1.5 million views. You can watch it below: