This Specialization is intended for all learners seeking to develop proficiency in statistics, Bayesian statistics, Bayesian inference, R programming, and much more. Through four complete courses (From Concept to Data Analysis; Techniques and Models; Mixture Models; Time Series Analysis) and a culminating project, you will cover Bayesian methods — such as conjugate models, MCMC, mixture models, and dynamic linear modeling — which will provide you with the skills necessary to perform analysis, engage in forecasting, and create statistical models using real-world data.
Amazing and capable professor, extremely interesting topic. I absolutely loved the fact that the coding was taught in both R and Excel and that the code was also available for the student.
The best course I had in statistics. unlike many other courses the instructor does not ignore the underlying mathematics of the codes.
Definitely quite mathematical in nature. Good way to learn about expectation-maximisation algorithm.
It was a nice course, but it would be better if there were more supplementary materials for the proof and theoretical discussion.
Amazing course ! Excellent. Prof. Herbert has a great didactics, the material is clear and very well planned, including the assessments. Thank you very much. Flavio Lichtenstein (Brazil).
The course was really interesting and the codes were easy to follow. Although I did take the previous course for this series, I still found it hard to grasp the concepts immediately.
I really enjoyed this course! Plenty of examples on how to use Mixture Models in a Machine Learning context. Thanks to Abel and his team for putting together such an useful course.
The concepts could have been made much more easier, simpler and more examples could have been inculcated. But overall, a great course to gain in-depth insights for a starter on Bayesian Statistics
Very good instructor, knowledgeable and thorough, touching the right level of details with big picture in mind, and providing practical guide for hands-on Bayesian data analysis.
I learned a lot about bayesian mixture model, expectation maximization, and MCMC algorithms and their use case in classification and clustering problems. I highly recommend this course.
It was a good course for me to get familiar with the new perspective on statistics. Thank you! Maybe, some extended practice exercise at the end of the course would make it even better)
A thorough and comprehensive overview of applied Bayesian modelling which will give you the confidence to start applying Bayesian tools in your own work.
Showing: 20 of 846
So, I really wanted to LOVE this class, but instead I found that I merely liked it, and want to use this review as a way to explain why. WHAT I LIKE ABOUT THE CLASS: The material is sufficient for the topic at hand, and is structured in an appropriate way. If you work through everything you'll have a decent grasp of exactly what the class is meant to be about. It's also pretty well paced. WHAT I DIDN'T LIKE ABOUT THE CLASS: Dr. Lee usually rushes through or skips discussions what concepts mean before formalizing them mathematically. As a result it's very easy to make progress through the class without a good feeling that you actually "get" what Bayesian statistics is really about. Too many of these videos are him chopping wood through the mathematical jingo, when the material DESPERATELY needed a 3-5 minute introductory video about what concepts actually mean or how to think about them. I remember telling my girlfriend during the middle of the class that I found it frustrating because I was progressing through it quickly, and getting the quizzes right, but lacked a good intuition for how to think about Bayesian statistics. So Dr. Lee......work on those presentation skills! Think deeply about how to communicate the essentials of the concepts in each lesson, and THEN start pounding away on the whiteboard!
I don't find that the lectures do a good job of relating the material to real world usage. To much focus on equations and too little on the why.
I felt like I just did a lot of calculations. The course was better in the beginning, as I felt the professor actually explained what and why were were doing what we were doing. By the middle of the course, however, I felt that the professor just jotted down equations and went really quickly. I don't actually understand why I was doing the calculations that I was doing.
Almost no intuition is given. I really got bored while watching the formulas to be written on the board without giving real meaning behind them. I would not have taken this course I was aware of these.
The instructor doesn't do a good job at teaching. He throws so many formulas at you without explaining any of them. The course is purely based on memorization not understanding the concepts. I have been using other online classes to be able to understand this class.
Herbert Lee is teaching by seeing books and write lots of equations doesn't explain how theory and equations related to real world applications. Its more like class room lessons , not like something that can be applied to real world scenarios.
If you already know everything about the topic and just forgot some little things or you are very strong in calculus, this may be a nice refresher. Otherwise, not very useful. Really dense and little explanation. I liked the Youtube MIT course on Probability (it includes Bayesian Statistics) much more, since it has good explanation of the concepts.
This course gives an introduction to the theoretical basics of Bayesian statistics. Before taking this class, I had a very confused view of the whole Frequentist vs Bayesian "debate". I understand now that Bayesian statistics is really about attaching uncertainties to beliefs and producing a clear definition of this uncertainty (especially through the notion of credible intervals).
The course really focusses on theory. I recommend knowing a bit of basic stats concepts before taking the class, such as Bayes' Theorem, basic discrete and continuous distributions, and confidence intervals. If you are not experienced with these, be aware that you will likely need to read-up on them throughout the course. R is used, but the usage is so simple that you should not shy away due to a lack of R experience.
I really have no complaints about the course. After completing it, you should understand the differences between Bayesian and Frequentist approaches. You will also understand a lot of terminology that gets thrown around in data science these days (priors, posteriors, credible intervals).
I get lost a bit too often.
The teacher sometimes explains easy concepts and omits the difficult ones (e.g. exponential distribution is explained as "for example if you are waiting for a bus that comes every ten minutes" and then he tells you how to compute expected value and moves on, but he does not say WHAT IT MEANS - is it the probability that I will meet an oncoming bus? is it probability of waiting ten minutes for the bus? is it the average waiting time? is it average number of buses that come every hour? - but there is detailed explanation of what A squared means in lesson two (!))
The teacher often makes me confused as to where he got the numbers he is plugging in the formula or what answer the formula gives.
But I take it as a challenge and I intend to finish the course despite all of that. Sometimes it is fun to decipher the mystic equations. And maybe it is me, maybe I was not born to be a statistician. Maybe there are people that find this stuff easy and understand it right away.
I really like the quizes. They are HARD.
One last thing: Wearing white shirt and using white marker makes it impossible to read what he writes. But I take it is part of the challenge ;-)
Good intro to Bayesian Statistics. Covers the basic concepts. Workload is reasonable and quizzes/exercises are helpful. Could include more exercises and additional backgroung/future reading materials.
I was baffled after the first lesson. There is no explanation or answers given.
A good MATHEMATICAL introduction to Bayesian Statistics. I read some of the negative reviews and they claim to have many formulas, well, that was exactly what I was looking because after watching some PyCon Videos about Bayesian Statistics I understood the code to solve the problem but not really why that code works or how.
This course may be frustrating for those with no prior introduction to Bayesian statistics, I recommend to take this course after seeing some videos from the Scipy, PyData and PyCon conferences regarding this topic.
This is a math course. It has good quality, it is rigorous and educational. It presents the mathematical framework of the Bayesian statistics. I like it when the math of the subject is explained well, as done in this course, rather than "I don't want to get in to the math", or "it is beyond the scope of this course", which you often see in online courses.
What I like about this course is that it goes into mathematical details and gives students with no or very limited prior knowledge about Bayesian statistics (even though some familiarity with general statistics is required) a pretty comprehensive introduction. Having almost completed the course, in retrospective I can say that I have learned a lot and also gained some intuition about Bayesian Statistics. However, here comes the problem: I think that I learned most when I resorted to textbooks or YouTube after not having understood the content provided by the instructor. The instructor does not try at all to provide intuition, he rather reads formulas from his notes and then writes them down. Of course, a topic this complicated requires that students do some additional work. However, Mr. Lee is definitely not a good teacher, and worse, he does not even try to be one – even a good teacher would not be able to convey a topic like exponential distribution and its usage in Bayesian statistics in just 4 minutes. In short: very good structure of the course, but pretty bad teaching style.
When doing the quiz for Lesson 2, Week 1, I first failed, then I used the hints provided to do the appropriate calculations, however the numbers I obtain are considered incorrect, and I cankot pass the quiz although I checked the calculations million times and I know I am correct. There should be a demonstration of how the correct answer is obtained, simpke hints are not enough. Although I was super excited about learning Bayesian statistics, now I am forced to quit as it looks I will never be able to complete the course. The course ended up being useless and frustrating. This is truly unfortunate
An excellent course with some good hands on exercises in both R and excel. Not for the faint of heart mathematically speaking, assumes a competent understanding of statistics and probability going in
Very useful to get an initial understanding of bayesian statistics. Recommends using R or excel but possible to use python using statsmodels.api, scipy, numpy and pandas tools
This is was a really difficult course. I took a basic statistics course in college but was not prepared for the calculus and the theoretical way this course was explained. If you are looking for a stats course that explains through real-world examples rather than theory - this ain't it. The only reason I gave it 2 and not a 1 star is because I can assume that those with a deeper statistical background would probably not face the challenges I did.
There should be more focus on understanding the equations. The R and excel videos were incredibly blurry.
This was a fantastic introduction to Bayesian statistics. Professor Lee is an excellent lecturer, with a comfortable, almost conversational style that I found easy to follow and stay focused on. The course itself is very well organized, introducing key concepts and then immediately providing examples that helped me internalize the concepts they pertained to. Quizzes were low pressure, straightforward applications of the lectures that served the purpose of allowing me to immediately apply what I had just learned.