Brandon - You must know, till now, whenever I have any exam - Statistics, Analytics, Operations, Accounting I always watch your videos just before the exam! Gives me a touchbase that revising from book doesn't provide! Please start posting advance videos on these topics.. Your videos are easy to understand and less daunting than so many other tutorials online.. Also, [i hope I am not becoming too demanding :)], can you also post videos on Corporate Finance and Economics (Micro and Macro) basics, from a Management student perspective? Would be grateful! Once again.. Thank you so much!
I need to point out an error on the MAPE example, the % error for years 2011,2010, and 2009 are correct, for every other year in the example, you started to divide by the forecasted value when you should really be dividing by the Actual value, |30,300-29,822|/29,822 should really be |30,300-29,822|/30,300.
I believe some of your calculations you made in the table are incorrect. For example, when calculating the the 3-YR MSE ERROR^2 for 2011: (30,530 - 30,476)^2 = 2916, where in your calculations it is 2947. I found similar mistakes in the calculations for all the other years except for 2009 when calculating the 3-YR MSE ERROR^2. My MSE = 60,407. Please correct me if I'm wrong. I just thought I should point this out for you. But thanks for your clear explanation of the methods - its very clear.
You could kinda arrive at each of these error measurements alone , even if you weren't introduced to it formally. They are simple and make sense, but are there any better measurements that are more complex and better for ranking types of forecasts? Or is it the best that we can do?
Thanks! just an observation, I think the MAPE values on the 4th column are wrong from 2005 to 2007... You've used the Forecast value to divide, and not the Enrolment value....
Sorry sir, I have some question about error measurement. I've got a 0 on my actual data, and it can't be measured specially on MAPE. Is there any solution for this problem? Thank you
I think MAPE is not relevant here as it is only used when there are many series. Also, using different measures can lead to different conclusions (when one method is better in terms of MAD and another method is better in terms of MAPE). Here's a paper saying why MAD (MAE) is better than MSE or MAPE: www.researchgate.net/publication/284947381_Forecast_Error_Measures_Critical_Review_and_Practical_Recommendations
Brandon - You must know, till now, whenever I have any exam - Statistics, Analytics, Operations, Accounting I always watch your videos just before the exam! Gives me a touchbase that revising from book doesn't provide!
Please start posting advance videos on these topics.. Your videos are easy to understand and less daunting than so many other tutorials online..
Also, [i hope I am not becoming too demanding :)], can you also post videos on Corporate Finance and Economics (Micro and Macro) basics, from a Management student perspective? Would be grateful!
Once again.. Thank you so much!
I need to point out an error on the MAPE example, the % error for years 2011,2010, and 2009 are correct, for every other year in the example, you started to divide by the forecasted value when you should really be dividing by the Actual value, |30,300-29,822|/29,822 should really be |30,300-29,822|/30,300.
Every Video is a treasure. Thanks
Well presented; clear, concise, articulate, informative.
Thanks a ton for these lessons..love the way you teach! These results match perfectly with my calculations including MAPE.
Great video Brandon, was a great refresher for a job interview :-)
I believe some of your calculations you made in the table are incorrect. For example, when calculating the the 3-YR MSE ERROR^2 for 2011: (30,530 - 30,476)^2 = 2916, where in your calculations it is 2947. I found similar mistakes in the calculations for all the other years except for 2009 when calculating the 3-YR MSE ERROR^2. My MSE = 60,407. Please correct me if I'm wrong. I just thought I should point this out for you. But thanks for your clear explanation of the methods - its very clear.
Exponential moving average was not explained in the previous video.
You could kinda arrive at each of these error measurements alone , even if you weren't introduced to it formally.
They are simple and make sense, but are there any better measurements that are more complex and better for ranking types of forecasts?
Or is it the best that we can do?
Thanks! just an observation, I think the MAPE values on the 4th column are wrong from 2005 to 2007... You've used the Forecast value to divide, and not the Enrolment value....
Sorry sir, I have some question about error measurement. I've got a 0 on my actual data, and it can't be measured specially on MAPE. Is there any solution for this problem? Thank you
You can try sMAPE, which I have a video for, or stick to MSE/RMSE. Zeros a menace! 😂
@@BrandonFoltz Thank you sir
Thank You
anyone can help me? how do we calculate the number of enrollment in the year 2005 (29,933). tnx
Thank ou Sir
Yellow text is difficult to read.
I think MAPE is not relevant here as it is only used when there are many series. Also, using different measures can lead to different conclusions (when one method is better in terms of MAD and another method is better in terms of MAPE). Here's a paper saying why MAD (MAE) is better than MSE or MAPE: www.researchgate.net/publication/284947381_Forecast_Error_Measures_Critical_Review_and_Practical_Recommendations