In the last few months, several people asked me this question “What are machine learning and data science?”
My usual crisp response before any explanation : It [machine learning and data science] is a lot of mathematics and a little bit of programming.
Indeed, explanation is required, especially to those who want to make a lucrative career in Machine learning and Data Science.
Firstly, to pursue any of these two paths you undoubtedly need at least 70% understanding of the following topics:
a) Topics in mathematics
- Probability : Combinatorics , Probability Rules & Axioms , Bayes’ Theorem, Random Variables, Variance and Expectation, Conditional and Joint Distributions , Moment Generating Functions, Maximum Likelihood Estimation (MLE), Prior and Posterior , Maximum a Posteriori Estimation (MAP) and Sampling Methods.
- Statistics : Measures of central tendency, Spread of the data and Standard Distributions (Bernoulli, Binomial, Multinomial, Uniform and Gaussian).
Multivariant Calculus: Differential and Integral Calculus, Partial Derivatives, Vector-Values Functions , Directional Gradient and Hessian, Jacobian, Laplacian and Lagrangian Distribution.
- Linear Algebra : pretty much everything under this topic.
Miscellaneous : some topics from here and there like . Information Theory and Game Theory.
b) Programming Language:
- Python is important for making career in Machine learning.
Few libraries and tools which will help you are pytorch,scikit learn,numpy,pandas,tensorflow and seaborn .
- R is For Data science as a career.
R is a better choice, but not enough. You are also required to learn few tools like : Tableau , Microsoft Power BI.
- MATLAB or Octave : For Research based field MATLAB and Octave make more sense. These tools allow you to test your hypothesis.
Data Science vs Machine Learning?
In terms of mathematics:
(there is a link available below this article for details.)
In terms of programming language:
It is all about which path you want to choose ML Engineer , Data Scientist or a Researcher.
How much time to invest?
The ideal requirement is 15 hours a week to learn. The following table as per your understanding level further categorise the required time to invest:
Links to some important -
- Introduction to Artificial Intelligence by Udacity.
- Machine learning on Coursera (Offered by Stanford)
- Mathematics behind Machine Learning — The Core Concepts you Need to Know
- Two Minute Papers