Moving Average processes - Stationary and Weakly Dependent

[@jonatanisse6362]

Man you're a life saver! This is simple and crystal clear! When my professor is discussing this topic it sounds like glossolalia.

[@martigonzalezferreres7093]

Ben, I do not have words to describe how useful are your videos... You are the best!!

[@bhaskarroy8753]

Excellent explanation Ben. Thank you very much.

[@shijingsi8288]

Thanks very much, Ben!! You are very helpful!!! So clear and easy to follow!

[@slippedgrey]

your videos are really wonderful for study econometric thank you for sharing!

[@bfindiy]

i let you are my professor on video, every chapter are extremely clear for understanding, really thank you for sharing

[@walcott0]

no wonder you have problem understanding the content

[@AnshumanKumar007]

4:57, this applies to the general proof for a moving average process where the individual rvs are not necessarily from a normal distribution

[@shaali4781]

Please upload videos of bayesian statistics

[@slippedgrey]

And I am studying time series econometric now do you have a video list for this topic? because I just don't know what to start with without a list

[@qderossi2790]

good video! It really helps a lot, thx!

[@antimoustique3175]

thank you for your video! just a question, does it mean that the MA(1) process is weakly stationary or strictly stationary ?

[@Djc99120]

Weakly until and unless the multivariate normality of xt s are assumed

[@stephg7404]

videos are great but wish they went to more depth, i get lost in the Variance and Covariance

[@realisttik]

Hello sir, is MA representation still a AR (1) process?

[@nickhollinger627]

Var(et+(theta)et-1) = Var(et) + (theta)^2Var(et-1) + 2(theta)Cov(et,et-1) correct? Are you not missing the 2(theta)Cov(et,et-1) when solving for Var(xt)?

[@nickhollinger627]

Or just eliminated b/c Cov(et,et-1) = 0?

[@lastua8562]

He explicitly said he leaves that out because of iid errors.

[@1vaas]

how can we arrive to the definition of epsilon(t) and epsilon(t-1)? if X(t) is calculated from e(t), how are we getting e(t) without knowing X(t)?

[@looyt]

+1vaas e(t) is a variable that exhibits a random behavior with mean 0 and var sigmasquared. examples of this variable might be change in temperature per day- it can take positive or negative values but the mean is 0. consequently, e(t-1) is the change in temperature yesterday.

[@1vaas]

+loo yuntong michael thanks for the answer. However, I understand that e(t) is a random variable with 0 mean and some variance. My doubt is how that error is occuring. suppose, e(t) is error at time point (t). that means it is difference between two values at time (t). what are those two values is my doubt

[@gymmel1991]

+Ashok Varma , Ashok, it's a shock! :)

[@shafiqnasran1]

Why is some of the MA(1) model that I have seen is different? For example, my lecture note states it as Y_t = e_t - theta * e _t-1

[@lastua8562]

I guess it is similarly an MA(1). It does not need to be positively related to the previous error, just related is enough. If you wish, you could also get a negative theta when estimating, making the whole thing positive.

[@revolutionarydefeatism]

It seems that after 7:00 there is a noise of football fans in the background. :-D

[@Tino51]

The ugliest sigma i've ever seen

[@lastua8562]

the most useless comment I've ever seen