Back to Mathematics for Machine Learning: PCA

4.0

stars

2,078 ratings

•

508 reviews

This intermediate-level course introduces the mathematical foundations to derive Principal Component Analysis (PCA), a fundamental dimensionality reduction technique. We'll cover some basic statistics of data sets, such as mean values and variances, we'll compute distances and angles between vectors using inner products and derive orthogonal projections of data onto lower-dimensional subspaces. Using all these tools, we'll then derive PCA as a method that minimizes the average squared reconstruction error between data points and their reconstruction.
At the end of this course, you'll be familiar with important mathematical concepts and you can implement PCA all by yourself. If you’re struggling, you'll find a set of jupyter notebooks that will allow you to explore properties of the techniques and walk you through what you need to do to get on track. If you are already an expert, this course may refresh some of your knowledge.
The lectures, examples and exercises require:
1. Some ability of abstract thinking
2. Good background in linear algebra (e.g., matrix and vector algebra, linear independence, basis)
3. Basic background in multivariate calculus (e.g., partial derivatives, basic optimization)
4. Basic knowledge in python programming and numpy
Disclaimer: This course is substantially more abstract and requires more programming than the other two courses of the specialization. However, this type of abstract thinking, algebraic manipulation and programming is necessary if you want to understand and develop machine learning algorithms....

JS

Jul 17, 2018

This is one hell of an inspiring course that demystified the difficult concepts and math behind PCA. Excellent instructors in imparting the these knowledge with easy-to-understand illustrations.

NS

Jun 19, 2020

Relatively tougher than previous two courses in the specialization. I'd suggest giving more time and being patient in pursuit of completing this course and understanding the concepts involved.

Filter by:

By Laszlo C

•Dec 06, 2019

This is an excellent course first covers statistics, looks back to inner products and projections, thereafter it connects all of that and introduces PCA. The knowledge that you've gathered throughout the first two courses gets applied here. Granted, it's more abstract and challenging than the others, I wouldn't give a worse rating just because of that. You'll need to dive into certain topics on your own and if you strengthen your coding skills for the programming exercises. Nevertheless, it's just as highly rewarding as the first two.

By Douglas W

•May 22, 2020

This was the most challenging of the three classes in the series. I thought the instructor did an excellent job of moving from theory to practice, and in the end I came away with a good understanding of the topic. As a developer, part of my personal learning style is to shadow these types of lectures in code. I did (or attempted) naive implementations on most slides - that definitely helped my comprehension of this challenging material. Be prepared to work hard, occasionally head scratch and you'll do fine.

By Jitesh J T

•Dec 24, 2019

Hi,

The course tries to cover most of the important mathematical concepts in Mathematics applied to PCA. The assignments were a bit tough, but i guess that the road ahead when we do programming for data sets in real world applications would not be that easy. Loved the way the lectures were delivered and the programming assignments help to build a strong base for applications of linear algebra that we have done earlier.

Thanks and Regards

Jitesh Tripathi, PhD in Applied Mathematics

By Tarek L

•Sep 11, 2019

This is a difficult course, but it really gave me an appreciation of the mathematics behind machine learning. I encourage anyone doing this course to read Deisenroth's free book Mathematics for Machine Learning (mml-book.com) to better understand the notation and technique used to get to the proofs. If anything, the rigor of this course inspired me to further pursue learning in mathematics to strengthen my machine learning foundation.

By ChristopherKing

•Apr 18, 2018

The whole content of this course is fantastic, not all details were covered in the video, but main ideas were expressed in a great way buy math formulations. Pay attention to those vectors and matrices, especially their dimensions, this will help you solve problem quickly. More important, matrix is just a way to express a bunch of similar things, knowing the meaning of those basis vectors is important.

By Sriram R

•Jun 18, 2019

This is one of toughest course in this specialization. Having said that, it was interesting to learn about the inner working of the PCA and is well taught. At times it was tough to follow and could have been better if there are some additional examples explained to reinforce the concept. Also week 4 is kind of rushed with little or no time to fully appreciate the beauty of PCA.

By Yuanfang

•Sep 07, 2019

A little more challenging than the other 2 courses in this series. The programming examples on K nearest neighbors, eigenvector fitting of facial data, and the PCA implementation are neat and rewarding. Can't help but feel there's still a great deal of math details that is only briefly mentioned - oh well there's always the free textbook to reference. Overall highly recommended.

By Marcelo R

•Jul 27, 2020

Unlike the other two modules, the course is quite challenging, some details are omitted in the explanation and one has to look for them in the forum or on the internet. Some notebooks for programming have problems and need to be downloaded and run virtually. Still, the content is exciting, thanks to the Imperial College London for the course and the opportunity.

By Renato

•May 03, 2020

This course is challenging, it requires a lot of participation in the forum plus an overlook on the internet to help you out understand a little more how the vector (eigenvectors) relate to the efficiency of PCA. It is pretty interesting to understand the algorithm itself and how it works. Be aware to review a lot and take your time to understand things.

By Gergo G

•May 15, 2019

This course is really challenging. A strong mathematical background is necessary or it needs to be developed during the lectures and self-study. The professor's explanations are clear, and still lead to complex ideas which is great. Programming assignments are also difficult, however they serve as a superb opportunity to develop your skills in Python.

By Anastasios P

•Dec 26, 2019

Challenging course, a lot harder than the two previous in the specialisation. Having said that, I really enjoyed it for the insights that it gave and for actually making me learn some Python as well. With this course you need to go search and fin the necessary functions and usage to complete the assignments. The best course in the series I believe.

By Idris R

•Nov 02, 2019

Great, challenging course. The instructor will expect much of you as the material is not spoon fed. At times this is frustrating but yet that's the best way to build your own intuition. This is a *hard* course and I imagine most of machine learning is like this. Fun, rewarding, and challenging. You'll flex your math and programming muscles.

By Jaiber J

•May 01, 2020

A great course, worth the money. It was hard, as it should be. The explanations are concise, and the assignments take much more to complete, at times leaving us scratching the head. Anyway, I'm so glad to have completed, it has provided me such great insight about how mathematics powers the machine learning algorithms we use everyday.

By Ratnakar

•Jul 13, 2018

This is by far the best course I have taken. The Instructor is exceptional in setting the stage to understand the complex topic by letting us know the motivation of every concept, making us understand the fundamentals right, deep diving into the core of the topic and them nicely summarizing the topic along with the applications.

By Geoffrey K

•Jun 05, 2020

This course is at a higher level than the first two in the specialisation, and the instructor focusses on the mathematics of matrices, while the assessments are programming. There are easier courses for just PCA (which I thought helped me). Looks like most learners find a way through, and its worth it.

By Fernando G M G

•Jun 30, 2020

It was a great course. Challenging at some points since I'm new in Python but it was worth the effort and I really learn a lot and now I comprehend the maths behind PCA algorithm. The point in which the relationship between eigenvalues of the covariance matrix is used in the PCA algorithm was amazing.

By Surbhi P

•Jun 17, 2020

Learning Mathematics in this way and in efficient manner from basics and very clearly is really nice. I am very thankful to this course , teachers, Imperial College London as well as team of Coursera for providing such a great platform to learn all these skills and enhance our knowledge.

By David L

•May 29, 2019

This was indeed a very challenging course. It was also very rewarding, and I felt that the instruction was great and relevant to the assigned tasks. The first two courses in the specialization were very high quality, and in my opinion this one lives up to the expectations that they set.

By FRANCK R S

•Jul 07, 2018

Very interesting and challenging subject: PSA, this MOOC together with the other 2 Mathematics for Machine Learning are one of the most useful I have ever made, actually they helped a lot in my other Machine learning and Deep learning studies! I highly recommend this fascinating MOOC

By mohit t

•May 13, 2018

Perfect course. It takes up more time and effort than the other two courses in the specialization. But what you learn by the end of it is totally worth the effort. Note that this is an Intermediate course compared to the other two which are beginner. So the extra rigor is expected.

By Oj S

•Jan 13, 2020

The introduction to PCA and steepest descent algorithms which might be a century old but still act the fundamentals of many state of art equations. So, you will learn the basics that how they function, and the real mathematics you need to know for ML using this course.

By anurag

•Apr 18, 2020

Its a very informational and interesting course. I understood a lot about PCA in this amazing course.

It was a good addition to the previous two courses of the certification. I would like to get similar courses in statistics and probability useful in Machine learning.

By Anna U

•Jan 14, 2020

An excellently simple explanation of concepts of linear algebra and PCA. Applause for lector. I really liked this course and found it very useful for those newbies in machine learning like myself. I recommend this course to all my friends and others interested in.

By Ramon M T

•Oct 23, 2019

I liked the course quite a bit. I found it quite challenging (I had never seen any PCA) but it always kept me very interested. I had to use several sources to read a little more about PCA and to complete the last exercises, the forum is very helpful.

By Bingfeng H

•Aug 26, 2020

Very good course, although the programming assignments are challenging and some background knowlege in linear algebra and vector calculus required. You will need to do some independent research at times. But the instructions are clear and concise.

- AI for Everyone
- Introduction to TensorFlow
- Neural Networks and Deep Learning
- Algorithms, Part 1
- Algorithms, Part 2
- Machine Learning
- Machine Learning with Python
- Machine Learning Using Sas Viya
- R Programming
- Intro to Programming with Matlab
- Data Analysis with Python
- AWS Fundamentals: Going Cloud Native
- Google Cloud Platform Fundamentals
- Site Reliability Engineering
- Speak English Professionally
- The Science of Well Being
- Learning How to Learn
- Financial Markets
- Hypothesis Testing in Public Health
- Foundations of Everyday Leadership

- Deep Learning
- Python for Everybody
- Data Science
- Applied Data Science with Python
- Business Foundations
- Architecting with Google Cloud Platform
- Data Engineering on Google Cloud Platform
- Excel to MySQL
- Advanced Machine Learning
- Mathematics for Machine Learning
- Self-Driving Cars
- Blockchain Revolution for the Enterprise
- Business Analytics
- Excel Skills for Business
- Digital Marketing
- Statistical Analysis with R for Public Health
- Fundamentals of Immunology
- Anatomy
- Managing Innovation and Design Thinking
- Foundations of Positive Psychology