Greetings, Ben. Your explanation is awesome! I have been looking for a intuitiveness behind the math for a really long time and i finally find it. In my opinion, there is nothing more valuable than getting the basic understanding of how to extract the final result starting from the beginning. You saved my time! I find this content extremely useful. Thank you!
Thank you so much! This helped me grasp a difficult concept for which no explanation was available in the book Have been watching your videos for a while, makes for crisp understanding of esoteric concepts, please continue your good work that I discover to be an art.
Thanks Ben! Just one question: why do we care about transforming between MA(1) and AR(infinity) as in this video (AR(1) and MA(infinity) in the previous video in the playlist)? What do we gain from transforming from one process to another? Maybe you said that, but I missed it
It isn't missing. I asked the exact same thing to my Professor the other day. The answer is that it IS NOT a moving average, it is a moving summation and so the title is a misnomer. Misleading af I say.
Hello, my teacher explained that not every MA process is invertible and that, for an MA(q) process to be invertible, the solutions of the polynomial equation have to be inside the unit circle. Do you know why is that? Thank you very much!
It just depends on the way that the data moves. A function can be defined however you want it theoretically, he has chosen a different starting equation.
+Friederike L Thanks for your comment. There are ways to do this using lag polynomials see: stat565.cwick.co.nz/lectures/06-ARMA-properties.pdf - it goes into some detail about how this might work. Best, Ben
Hi everyone, can someone help me to understand what is the phrase that he mentioned in minute 1:11, I understand sometihing like "the lang of.." Thank u so much
![ARMA Stationarity, Invertibility, and Causality [Time Series]](http://i.ytimg.com/vi/uciHswYSA3k/mqdefault.jpg)
![เจ็บที่ยังฮัก - ม่อน วรวิทย์ [ Official Mv ] จอนนี่มิวสิค](http://i.ytimg.com/vi/KALIGSRSjFs/mqdefault.jpg)
Greetings, Ben.
Your explanation is awesome! I have been looking for a intuitiveness behind the math for a really long time and i finally find it. In my opinion, there is nothing more valuable than getting the basic understanding of how to extract the final result starting from the beginning. You saved my time! I find this content extremely useful. Thank you!
Thank you so much! This helped me grasp a difficult concept for which no explanation was available in the book
Have been watching your videos for a while, makes for crisp understanding of esoteric concepts, please continue your good work that I discover to be an art.
Thanks a lot! This helped me get a full mark on a 20 mark final exam question!
Hi, glad to hear it helped! Best, Ben
thanks for your work, you saved someone today!
Ben you are a ledgend
Ben please help me how to interpret the ADK test
thank you so much, some just explain it without mentioning the steps in between, as we are supposed to know these steps
Very good explanation!
I really enjoyed listening to this lecture
straight to the point. Thank you so much
Hello I am just wondering isn't MA(1) process defined as:
xt = et + theta * et-1 instead of et - theta * et-1?
theta can be positive or negative , depends upon how you want to model .
Both possible. Just switches the sign on theta.
Thank you. This was terribly helpful.
Thanks Ben! Just one question: why do we care about transforming between MA(1) and AR(infinity) as in this video (AR(1) and MA(infinity) in the previous video in the playlist)? What do we gain from transforming from one process to another? Maybe you said that, but I missed it
We want to show the similarity, just that they are not all that different from one another but rather related.
Very useful! Thank you very much!
Very succinct thought
Your MA(1) model is missing the mean in the equation, how does a sum of the errors equal the actual estimate of the value at time t?
It isn't missing. I asked the exact same thing to my Professor the other day. The answer is that it IS NOT a moving average, it is a moving summation and so the title is a misnomer. Misleading af I say.
Hello, my teacher explained that not every MA process is invertible and that, for an MA(q) process to be invertible, the solutions of the polynomial equation have to be inside the unit circle. Do you know why is that? Thank you very much!
Because the geometric series would not converge otherwise
my lecture used Yt = Epsilont + tita x Epsilont-1 may I ask why the equation in my lecture uses + but the example in the video uses minus?
It just depends on the way that the data moves. A function can be defined however you want it theoretically, he has chosen a different starting equation.
How would it look like if we had an MA(2)process?
Thank you so much!!!!!
How does it work for higher order MA(q)?
+Friederike L Thanks for your comment. There are ways to do this using lag polynomials see: stat565.cwick.co.nz/lectures/06-ARMA-properties.pdf - it goes into some detail about how this might work. Best, Ben
Can an AR(1) be converted to an MA(infinity) ?
Yep, via the Koyck transformation: th-cam.com/video/hppHTnz0NgU/w-d-xo.html Cheers, Ben
0:04 Sorry how do you spell that? Coike transformation? Thanks
Koyck
Literally same
Hey would it be the same if an intercept is included?
i have the same question
Hi everyone, can someone help me to understand what is the phrase that he mentioned in minute 1:11, I understand sometihing like "the lang of.."
Thank u so much
The lag operator.
THANK YOU! :)
Thalk you!
Does anybody know how to convert an arma(1,1) process to an ar(infinity)?
You haven't started with an MA(1) though. You've started with an AR(1). Davidson and McKinnon 273
Hey did start with an MA(1) because we have theta, not phi.
Great