Brandon is perhaps the finest statistics instructor on youtube....his content can be easily internalised by people of non math / statistics backgrounds. You are doing a fine job there Brandon! May the force be with you.
I love the videos. At minute 21 I get for 2005 calculation [ 29025 * 1 + 29860 * 2 + 29953 * 4 ] / 7 = 29,793.86. Funny seeing the WMA for 2006 of 29,928.29. Funny that five years have gone by and no one commented on the math or the year misunderstanding. The videos are helpful, thanks!
Hello and thank you for your message. Pro-level forecasting is incredibly complex (which is why financial firms hire math Ph.D.'s! ) :) They are complex software packages. If you have other stats training, time-series regression is something you can look into. The best case studies are on the web and here on UA-cam. I use case studies in my books. I also have my students do their own; pick a stock, company, weather pattern, enrollment pattern, etc. and then do some forecasts.
Hello! Thank you for your comment. Ah yes...I do believe I had a transpose error from Excel into PowerPoint. The WMA at 20:43 should be 29,794. Thank you for pointing that out. I will add an annotation shortly. All the best, B.
i was going to comment that i got a little confused with the wma's because i keep getting different figures. :) thank you for uploading these videos! makes my life much easier. very simple and easily understood! thanks for the lecture sir! :)
Great explanation. Regarding WMA results, as mentioned by @Charles Schwer , 2005 was miscalculated. When corrected, this new serie also lags the original data. So what's the catch here? Thanks by the way.
The weightings I used in this example are arbitrary. There are methods for trying different weights and then testing them against historical data to see which is most accurate. It really depends on the data. If recent history is more important then recent data is weighted more. So there is no hard-and-fast universal rules with respect to weighting.
hi brandon.. i have been following your videos since last one year. i have one request.. can u make time series models like ARMA , ARIMA , exponential smoothing in R thanks
hi, thnx fr ur video but can u tell me where i can learn more about forecasting at professional level except college and books..just gve me links of case study / Pdf or anything else ..
Brandon is perhaps the finest statistics instructor on youtube....his content can be easily internalised by people of non math / statistics backgrounds. You are doing a fine job there Brandon! May the force be with you.
Many thanks from Iran, you should be a role model for all educators.
Brandon, you're simply "the man"! You're getting me through my job on a daily basis. Thank you and God Bless you!
I love the videos.
At minute 21 I get for 2005 calculation [ 29025 * 1 + 29860 * 2 + 29953 * 4 ] / 7 = 29,793.86. Funny seeing the WMA for 2006 of 29,928.29. Funny that five years have gone by and no one commented on the math or the year misunderstanding. The videos are helpful, thanks!
Hello and thank you for your message. Pro-level forecasting is incredibly complex (which is why financial firms hire math Ph.D.'s! ) :) They are complex software packages. If you have other stats training, time-series regression is something you can look into. The best case studies are on the web and here on UA-cam. I use case studies in my books. I also have my students do their own; pick a stock, company, weather pattern, enrollment pattern, etc. and then do some forecasts.
Best video so far. I am waiting for ARIMA.
Hello! Thank you for your comment. Ah yes...I do believe I had a transpose error from Excel into PowerPoint. The WMA at 20:43 should be 29,794. Thank you for pointing that out. I will add an annotation shortly. All the best, B.
i was going to comment that i got a little confused with the wma's because i keep getting different figures. :)
thank you for uploading these videos! makes my life much easier. very simple and easily understood! thanks for the lecture sir! :)
Great explanation. Regarding WMA results, as mentioned by @Charles Schwer
, 2005 was miscalculated. When corrected, this new serie also lags the original data. So what's the catch here? Thanks by the way.
Thank you Brandon!!!! this video has really help me
Great videos!!! You the savior of the day! i love your jobs!
watching from Brazil, thanks!!!!
Brandon.. Fabulous .. can you please create videos on other time series concepts.. like AR process, cross correlation etc?
The weightings I used in this example are arbitrary. There are methods for trying different weights and then testing them against historical data to see which is most accurate. It really depends on the data. If recent history is more important then recent data is weighted more. So there is no hard-and-fast universal rules with respect to weighting.
hi brandon.. i have been following your videos since last one year.
i have one request.. can u make time series models like ARMA , ARIMA , exponential smoothing in R
thanks
I cannot thank you enough Brandon......
Thank you Brandon!!!
Hi Brandon,
Thanks for the video.Can u help me with how to assign weights to our periods??
hi .. i wanted to knw how to forecast using moving average for 2012 to 2016 using ur example using ur second method
Thanks
Thanks man :)
hi,
thnx fr ur video but can u tell me where i can learn more about forecasting at professional level except college and books..just gve me links of case study / Pdf or anything else ..
Did you ever find out how close the fall 2012 enrollment prediction was to the actual enrollment in 2012 and which method was closest?
Giving there is a trend it probably under estimated it.
Moving averages are good when there is no trend of the value you are trying to calculate.