How to learn Machine Learning ?

STEP 1: Learn the Basics:

Get the knowledge of Machine Learning in theory, . Understand the importance of Statistics and its use.
Learn the mathematical concepts from usage perspective (Read: Maths as a Language) and not just as a subject with numbers and graphs.
Learn or Revise programming language of your choice. It could be Python, R Language or Java.

STEP 2: Decide Algorithm

One of the effective ways to learn Machine Learning is to decide an algorithm and work on it till you understand it deeply. Various courses teach algorithms one after the other. They have to finish the course within timelines. If you learn lot of algorithms without understanding, then it will of little use.

While learning any Machine Learning Algorithm, keep the below points in mind.

• Exact Name of Algorithm.
• Type of problem it solves.
• Which situation is best to use the algorithm.
• Which situation is not suitable to use the algorithm.
• How to use and what inputs are needed.
• How to evaluate
• How to improve its efficiency

Solving some real life examples available on the internet is one of the best ways to grasp and understand.
Also as far as possible have considerate timelines to finish a single algorithm which may range from a week to a month. Don’t try to learn everything in one go.
Have patience. Once you get the crux of the algorithm, go to next algorithm.

STEP 3: DOMAIN MATTERS

Remember Data Science has three legs as we saw in Basics of Data Science.
The technical, the mathematical and the domain. The domain leg is many times left to SMEs (Subject matter experts) of that particular domain. A data scientist should know that his/her most important duty is to add value to business. For this, understanding of business domain is very essential. The more you know the domain, the better are your chances to add value to business along with the models, suggestions and other outcomes.