I like all your videos Mr. Eugene. Accept my all respect and admire.

Remarkable as usual, Mr. Eugene. I like all your videos.

Thanks so much! this was much easier to understand than the other videos I watched about this. Thank you for not rushing through the equations like others do, assuming everyone's had 10+ years of Excel experience.

this guy is wonderful. keep up, Eugene

Thank you for posting this sir!

Thank you so much for explaining so well.

Very helpful, thank you!

Thank you so much for the help!!

Outstanding..Thank you for free education..

this video is very helpful and very clear.however, i am wondering whether you have tutorials on EWMA ?

Wow you are a life saver. Thank you

Top draw explanation and very helpful. Thank you!!

Helpful! Thanks from a industrial engineer

Perfectly explainede, this video solves my concept.

thank u sir , i have an econometrics lab test coming ur video helped me understand so much

I get it, thanks so much!

thanks for walking through the formula

Excellent professor

Thanks is very helpful

Sir I am really thankful for your tutorial and guide , It really helps me 💕🤗

@snehaltheprogrammer

2 жыл бұрын

How can he get Alfa and intial forecast can you tell me please 🥺

Thank you!

You are great sir❤️

Sir, thank you so much.

How do you find week 16 forecast? You have ran out of actual values so you cannot use the formula. Trying to understand how to forecast several periods into the future.

@KrampuZ_Jhonnys6508

2 жыл бұрын

I'm in the same situation. have u got any lucky with that?!

@megane.lthomas2066

2 жыл бұрын

@@KrampuZ_Jhonnys6508 he replied elsewhere talking about Bootstrapping… essentially I think for the Brown method, since the demand is without trend or season. We can therefore anticipate several periods to come with the same figure.

Great explanation

Thanks a lot sir

Eugene, Is there away to include lead time?.

Hello! just to confirm, when you say previous time series, he is saying all previous data points not just the one above. This is why it is called exponential smoothing. Wherein FT-1 is the result of the previous FT-2, and FT-2 is result of FT-3, and so on.. so this is a exponential model in the long run.. You can also add Trend and seasonality to the equation.

thank you!!

@EugeneOLoughlin

Жыл бұрын

Welcome!

I need help. I'm giving an x and y data set on is a series of years, 1993, 1994, 1995,...2007, the second data set is various decimals. 2.08, 2.24, 2.34... How would I compute this wit hout knowing alpha and the initial forecast?

Hello, I got two questions Mr. O' Louglin: 1. How did you calculate the smoothing constant to 0,6? 2. How was the value in B19 calculated, is B19 the average of the cell B? Thank you!

@EugeneOLoughlin

4 жыл бұрын

Hi Ano, I did not calculate the Smoothing constant - this is decided by you, the researcher. It is a number between 0 and 1.0. Here's some materail from my course notes about this: What value should I use for the smoothing constant? The smoothing constant determines the level at which previous observations influence the forecast. Large weights result in faster changes in the fitted line; small weights result in slower changes in the fitted line. Therefore, the larger the weights the more the smoothed values follow the data; the smaller the weights the smoother the pattern in the smoothed values. Thus, small weights are usually recommended for a series with a high noise level around the signal pattern. Large weights are usually recommended for a series with a small noise level around the pattern. A different way of choosing the smoothing constant: for each value of α, a set of forecasts is generated using the appropriate smoothing procedure. These forecasts are compared with the actual observations in the time series and the value of a that gives the smallest sum of squared forecast errors is chosen. (The commonly used values for a are between 0 and 1, with steps of 0.2 or less.) Hope this helps, Dr E.

Thanks you sir

life saver

thank you for the helpful video. Sir, I wanted to know that is it okay if I place the value of alpha as "0.08"?

@EugeneOLoughlin

8 жыл бұрын

+Arslan Zaidi I recommend that you check out the following web page "What value should I use for the smoothing constant?" from Minitab at: support.minitab.com/en-us/minitab/17/topic-library/modeling-statistics/time-series/time-series-models/what-value-should-i-use-for-the-smoothing-constant Dr E.

@sayanchakraborty6589

8 жыл бұрын

Actually you can use solver optimization technique

Wow thanks

Couple of questions- how did you decide on 0.6 as alpha value and how 65 was considered as the initial forecast

@NuM83RN1N3

9 жыл бұрын

Shivi Bhatia i'd like to know it too

@EugeneOLoughlin

9 жыл бұрын

NuM83RN1N3 and +Shivi Bhatia I recommend that you check out the following web page "What value should I use for the smoothing constant?" from Minitab at: support.minitab.com/en-us/minitab/17/topic-library/modeling-statistics/time-series/time-series-models/what-value-should-i-use-for-the-smoothing-constant

@oliviaee7812

5 жыл бұрын

​@@EugeneOLoughlin the website is invalid

@buddhaspeaks9723

5 жыл бұрын

@@EugeneOLoughlin thank you This is really easy

thank you

im supposed to make a forecast for 24 periods, so i don't have the previous periods actual demand for 23 of those periods. how am i supposed to calculate the others?

Tnq u sir

This is a helpful video, but your forecasted values via smoothing consistently underestimate the Actual Demand, is this an issue with alpha?

@EugeneOLoughlin

9 жыл бұрын

Hi Evan Freethy - I recommend that you check out the following web page "What value should I use for the smoothing constant?" from Minitab at: support.minitab.com/en-us/minitab/17/topic-library/modeling-statistics/time-series/time-series-models/what-value-should-i-use-for-the-smoothing-constant

This is a great video! One question, how do you calculate week 17?

@EugeneOLoughlin

Жыл бұрын

Hi concert, I have forecasted the demand for week 15 in the video. When the actual demand for week 15 is known, this should be added to the table. Then you can calculate the forecast for week 16. When the actual demand for week 16 is known, this should be added to the table. Then you can calculate the forecast for week 17. This is a rolling forecast technique which in my example you can only do on a weekly basis to forecast the following week. I hope this makes sense. Dr E.

@abdullahhasan779

Жыл бұрын

@@EugeneOLoughlin Hi, If we are not using the rolling forecast method, and are using only exponential smoothing to predict the future data, how will that work exactly? We used your video example to calculate for just one additional time period, but what if we wish to calculate for further periods? what would be the approach to achieve that?

@EugeneOLoughlin

Жыл бұрын

Hi Abdullah, As described in my previous comment, this technique is only useful for calculating the next period (a week in my case). Each calculation relies on knowing the previous Actual value (A t-1) and previous Forecasted value (F t-1). Say, in my example at week 15, you wanted to forecast for week 20. Then you would have to know the actual value for 19, which you can only determine at the end of week 19, and the forecasted value for week 19, which you can only calculate in week 18. I hope you can see that without figures for weeks 16, 17, 18, and 19 - it is not possible (using this technique) to accurately forecast week 20. You could of course substitute values (eg using a simple or a weighted moving average) for Actual values with which you can then calculate forecasted values - but of course accuracy will be lower. If you are a College student you should check out books like "Operations Management" by Professor Nigel Slack - this has an excellent explanation about how Exponential Smoothing works. Hope this helps, Dr E.

How to calculate for week 16 or 17, whats is the previous actual demand in that case?

@WellingtonSantos-ue5po

5 жыл бұрын

You can use bootstrapping technique. Use exactly the same formula but you "freeze" the last actual observation (so called y at "origin").

how do you find smoothing constant

How do I forecast for week 16? I only have the forecast value for week 15 and not the actual value.

@WellingtonSantos-ue5po

5 жыл бұрын

You can use bootstrapping technique. Use exactly the same formula but you "freeze" the last actual observation (so called y at "origin").

Keep in advance Sir

How you find value of alfa can u explain

Tx sir

Can the alpha equal to 0?

very useful, is there a specific way you have calculated the initial forecast for week 1? If it isn't given to me in my exam, what should I use?

@EugeneOLoughlin

Жыл бұрын

Hi Reece, No - it is not a calculated value in this case. There are a few options that you could use. If you have an actual for week 1 you could use that. Also, if there is a final calculated value for a previous estimate, you could use that. You can also use simpler techniques, eg, a Moving Average, to calculate a starting value. Hope this helps, Dr E.

Can I use this method to forecast income? Or does it work just for short periods of time, weeks, months? I need to forecast the income for a few years.

@EugeneOLoughlin

4 жыл бұрын

Hi jam jam, This is mostly used for short-range rolling forecasts. Dr E.

@jamjam0409

4 жыл бұрын

@@EugeneOLoughlin Ok. Thank you.

Any one can help me , How to measure alfa

can we predict future data using the exponential smoothing? sorry iam new here

@EugeneOLoughlin

2 ай бұрын

Hi @thed4404, Exponential Smoothing is all about making forecasts - it is a really good method. It is great for forecasting for the next period (eg next month)- not so good for the period after that and beyond (neg the month after next). Study the video carefully and hopefully you will see what ES is all about. Hope this helps, Dr E.

How to get the initial forecast?

@sowmyak2447

4 жыл бұрын

Hey Carces, for the initial forecast you can take the average of the given Actual forecast(i.e, 74.00). That will be a good start rather than assuming any random number.

this is Single Exponential Smoothing

To improve your alpha: 1. Take the difference between Columb B and Column C 2. Sum the differences at cell D18 3. Run the goal seek function, setting D18 to zero

Sir, Greetings! Please, tell me how to calculate initial forecast value if you don't have any data.

@EugeneOLoughlin

Жыл бұрын

Hi S D, The initial forecast is not normally a calculated one - there are a number of ways to estimate what should be used. The first (and simplest) way is to use the Actual demand for week 1 - in the video above this would be a value of 58. You could also use a figure from a previous forecast. For example, if Week 1 was the first week of January, you could use a figure from the last week of December in the previous year. You could also use other methods to calculate a value. For example, a simple moving average (as I have done above for weeks 2, 3 and 5). Do note that the initial value affects all resulting forecasted values, you can adjust this value as you see fit. Hope this helps, Dr E.

@SD-un3ii

Жыл бұрын

@@EugeneOLoughlin Thank you so much.

But the trend is linear and not exponential

why 0.6 to alpha

why not 0.9 or 0.3