I watched this video to study for my stats 2 final... I was so happy to hear some very kind words of encouragement at the beginning. Teachers like you who care about their students (or in your case, just all students in general) are a treasure for this world. Thanks for being part of that amazing group of people who make the world a better place through caring education.
I'm only one minute into the video and I already have to thank you for this. So many people get down on themselves for not understanding a concept without realizing just how far they've come. Thank you for being so positive! We need more people like you.
Your pep talks, and I don't use that word to trivialise your intentions, are very encouraging. They have the quality of being uplifting by not being condescending. My ambition is to be a qualified teacher in the next 2-3 years and I really, really hope I can learn to express things as clearly, and positively, as you do sir.
@resal81 Hello! Thank you for watching and for your question. Covariance, correlation, and regression are all based on a matching concept for each variable. For example in this video, the returns are "paired" by month. If we were doing a height/weight analysis for people, the height/weight pair would be for the same person. The pair order does not matter so long as they remain a pair. Hope that helps! All the best, B.
Hi, As a worker in the field, who has been out of statistical analysis for about 15 years, I found you videos (and I've watched over 20 of them) very clear, concise, and helpful. Your videos are much clearer than most of my statistics books. Thank-you very much for the time and effort to make these videos and making the statistical analysis so clear.
How long has it been since I heard such encouragement! "Saruman believes it is only great power that can hold evil in check, but that is not what I have found. I found it is the small everyday deeds of ordinary folk that keep the darkness at bay… small acts of kindness and love."
Hi, The one thing that impressed me most is the quotes that you have used in every video of yours! Needless to say, I am understanding Stats better from watching your videos. Thanks so much.
I am working on a self-balancing robot that implements a Kalman Filter. It uses covariance as part of the calculation, and I spent about 5 days trying to learn it to no avail. I completely get it now. I just wish I had found your videos sooner. Thank you!
Thank you so much for your encouraging words and for making this confusing subject easier. No one takes the time to explain like you do.. you are truly a teacher. Thanks again
Thank you for actually caring about teaching and education. It makes all the difference and I've forwarded and shared these videos with my classfellows.
I literally FAILED multivariate statistics and these videos are making everything clear in preparation for my doctoral COMPS exam! Bradon, simply saying thank you will not suffice. The English language needs another word for your sacrifice and expertise!
Thanks a lot 🙏... I just started hearing about all these in my Data mining class and most of my peers seem to know about these and I started freaking out. Your videos are very descriptive and crystal clear. 🙂
This is really a good one. You have explained like a primary school teacher which is what beginners want. To have more clarity, you may show the data set that was used, also before going for the computations. In fact, covariance matrix is obtained by finding the variances between each variable and then grouping it as a matrix. What is the significance of this matrix as compared to looking/analyzing individual covariances? Suppose the variables are height, weight, wealth and education of 20 people, how do we explain the physical meaning of the covariance matrix obtained from this data?
All I can say is Wow!! Ok. Let me add these: I'm currently taking a class on controls and state estimation, where I'm learning about Kalman and Particle filters. The covariance matrix part was not clear to me until now! Thank you!
Constructive comment: you do not need to repeat the upfront message. It's nice, but only needs repeating once or twice. If somene persists and follows the later videos in your series then just tell them to refer to the first intro video for this motivaitonal stuff. Don't underestimate your followers for their ability to be self-motivated. They wouldn't be learning via RU-vid if they were not already somewhat motivated. Having said this... you do a better job than I can in teaching via online vids. So keep up the great work Brandon.
Hey Brandon! Big fan of all your videos and content for years! Your explanations have been a backbone to my stats learning that I started using during my undergraduate degree and continue to do it till this very day when I am now doing my PhD! Can I request you to make a video on multiple linear modelling. Thanks, Love Akira
Thank you for this video, Brandon. I'm taking an online class on robotics, and the instructor just skipped over explaining covariance matrices (whilst teaching something called Kalman Filters). This video helped a lot!
You mention that covariance values only tell us positive or negative relationship unless it is "at or around zero". Can you give an approximate number range of just how close to zero the number must be to tell us that there is probably no relationship?
Thank you for a clear and easily understood lesson on covariance. request you to teach factor analysis and principal component analysis as well, we will all greatly benefit from such a lesson. Thanks.
Thank you so much for this incredible video ! it was so helpful ! I've never heard such clear an explanation (and the intro was exactly what I needed), so thank you again !
Very nicely explained. One comment though. In the MS Excel correction factor you mentioned N/n-1. Although its obviously N/N-1, but makes me think why N & n.
Very excellent video teaching here. I do not expect learning statistics can be so easy especially the conceptual aspects. Very nice work. For me, if higher level topics could be covered, I will be much appreciating such as logistic regression, count model, odd ratios, chi-square, ...
I am struggling to get to know Structural equation modeling. I starts to read book and watch RU-vid video for understanding it, however there are many unfamiliar terms for me. Do you have any advices for me to understand SEM easily? i am not familiar with statistics before except for the regression (i learn this from you when i did my research using regression). However, things are getting tough now, i have to get to know SEM for my research. I am struggling indeed. I know it is very time costing to make a video as i am myself a youtube creator as well, so i am very appreciated if you could give me any hint for this? I can get a lot of sources for SEM, but you are the best who can simplify everything.
Wouldn’t it be better to say that the variance at the diagonal measures the variance of the variable with respect to the sample as a whole, rather than to say that the diagonal expresses the variance of the variable “with itself”? Any way, excellent videos, thank you very much! They are enormously helpful.
MS Excel 2010 provides covariance.p and covariance.s. It provides covar only for compatibility Also, your example overstates the population/sample problem. In real world situations, N is usually large enough that N or N-1 should make no difference at all to within working precision.
Hello, great job. Don't you stop doing these amazing videos. I have a question : what is your method for analyzing correlation between qualitative and quantitative variables ?
you should rather create a single video of about 2 hours rather than repeating the last vedio for almost half of the time.. but these helped me a lot .. Thanks
sir this video really help me a lot, but i have one question, why the sum of all the elements of the covariance matrix rather then the sum of all elemens on the diagonal is the variance of the sum of the random variables involved? thank you