Forecasting: Weighted Moving Averages, MAD

Поділитися
Вставка
  • Опубліковано 28 гру 2024

КОМЕНТАРІ • 240

  • @shadowhalk225
    @shadowhalk225 4 роки тому +76

    Sir, I'm not exaggerating when I say you've saved my life. God bless you

    • @kingchisale4605
      @kingchisale4605 2 роки тому

      Ndipo sukunama😂

    • @dafulegend5905
      @dafulegend5905 Рік тому

      I know this is old, but like because of this channel, I was basically able to get 3.0 with a terror prof. For more context, the professor grades infamously low, but I still managed to get a score of 3.0, which is around 85%, or a B - B+.

    • @philippalaribambilla8439
      @philippalaribambilla8439 Рік тому

      Thank you sir, this is very interesting

  • @mehdirahan6943
    @mehdirahan6943 3 роки тому +45

    If only professors could be as clear as this video as well as Joshua's other videos. Thank you so much sir!

    • @ivageorgieva2985
      @ivageorgieva2985 Рік тому +1

      Rightttt our professor literally makes it sound so much more difficult ..explaining such a simple thing for over 30 min when it can be shown for 5 minutes ..

    • @mehdirahan6943
      @mehdirahan6943 Рік тому

      @@ivageorgieva2985 It's also why I believe actual experience is worth way more than knowledge for the most part.

  • @alehegnasmie6628
    @alehegnasmie6628 2 роки тому +7

    Sir, I'm not exaggerating when I say you've saved my life. God bless you You Give me So much better than my teacher's explanation. Thanks! a lot

    • @joshemman
      @joshemman  2 роки тому +2

      Glad to help Alehegn.

  • @Skiiandme
    @Skiiandme 3 роки тому +5

    Saved me 5 hours of reading a complicated book. Thanks!

  • @apple_yt7122
    @apple_yt7122 7 місяців тому +2

    Thank you. You literally explained something that my professor couldn't for 2 straight hours. Thanks a lot.

  • @issact5932
    @issact5932 7 років тому +2

    You are a better Professor than MY professor !!!! more clear and easy to understand!!! 100% GOOD for international students!

  • @connornicholas8628
    @connornicholas8628 4 роки тому +6

    You have saved me! I've been stuck on a problem like this for hours!

  • @fadzreenamyra7919
    @fadzreenamyra7919 3 роки тому +2

    thank you sir, it took me 5 min to understand it clearly, u deserve everything in this world 😭🙌🏻

  • @krisramirez7139
    @krisramirez7139 5 років тому +8

    Thank you! I was so confused by this in my textbook. You broke it down to where I can understand the material!

    • @kanhaythakore8136
      @kanhaythakore8136 2 роки тому

      lode ka textbook mai galat diya hai
      sab loda bhenchod bakwaas hai

  • @immaculatechelangat2729
    @immaculatechelangat2729 2 роки тому +3

    Now I can pass my exam! Phew... Thank you for this. Very clear

  • @alperari9496
    @alperari9496 Місяць тому

    Coming from Deep Learning class where we use this method during the optimization of trainer. Thanks a lot.

  • @PhrAntoine
    @PhrAntoine 4 роки тому +11

    The 4 weeks percentages were already convent to a percentage, that's why it adds up to 1.
    That's also why you divide the WMA by 5 for week 2.
    3(x)/5 + 2(y)/5 = (3x+2y)/5
    I think the important part to remember is to convert to percentages

  • @syedabrar9431
    @syedabrar9431 Рік тому +1

    Out class video... My exam will be conducted in a few hours I'm seeing this video all concept cleared. Thank you. God bless you dear. More n more shared in my class mates Whatsapp group

  • @P_sh9
    @P_sh9 Рік тому +1

    Legit needed this vid. Thank you. If it wasn't for this vid I'd be stressed as hell.

  • @ShinEKMG
    @ShinEKMG 2 роки тому +1

    Very helpful video. This channel must be located by students.

  • @md.yousufalamin6348
    @md.yousufalamin6348 3 роки тому

    Thank you man tomorrow is my exam. It helped me. Great video.

  • @muhammadtahakhan8388
    @muhammadtahakhan8388 6 років тому +2

    Given the following data, use exponential smoothing with α = 0.2 to develop a demand forecast for period 7. (Forecast for Period 1=10).

  • @shaikhashab9697
    @shaikhashab9697 2 роки тому

    you are a great teacher sir everyone is a teacher but everyone is not have a Idea how to teach you teach very well sir

  • @pagalhokya
    @pagalhokya 2 роки тому +2

    You're a legend! I wish my professor was like this :(

  • @adnandanisubhanallah8218
    @adnandanisubhanallah8218 2 роки тому

    i just fall in love with your method of explaining🙈🙈🙈

    • @joshemman
      @joshemman  2 роки тому

      Thank you so much 😀

  • @twylahjanet
    @twylahjanet 5 місяців тому

    Awesome job of breaking it down!

  • @triciaeisaac
    @triciaeisaac 4 роки тому

    Thank you sir! Your videos are clear and I always leave smarter than when I first arrived!

  • @kaelynf12
    @kaelynf12 3 роки тому

    this was very helpful before I had to take an exam. Thank you!

  • @margaretrajski4808
    @margaretrajski4808 Рік тому +1

    Excellent explanation!

  • @PhrAntoine
    @PhrAntoine 4 роки тому

    I would suggest changing the values for weeks to a date structure. It makes it easier to follow. The term "next most recent" is an odd way to say the previous week

  • @HealthyFoodBae_
    @HealthyFoodBae_ 3 роки тому +2

    Thank you!
    1. Is weighted moving average, or moving average really a forecasting method? Or is it a way of smoothing the data?
    (I am trying to learn and understand. )
    2. How do you determine which weight value to use?
    Thank you for the insight.

    • @joshemman
      @joshemman  3 роки тому +1

      1. They are smoothing methods and can be used for forecasting.
      2. Typically you choose your weights in a way that minimizes error. Methods used will vary by context.

  • @namiqqfrli3553
    @namiqqfrli3553 8 років тому +1

    Man,
    during 3:42, I think that something went wrong. So, for week 5 we should have to calculate (3*45+2*40)/5. 2*40 not 44. Because as seen all other calculations we have done it the first 2 CONSEQUENCE numbers from back to front.

    • @joshemman
      @joshemman  8 років тому +1

      You're right. That was a typo. The answer is correct though.

  • @tharmasilan.v335
    @tharmasilan.v335 2 роки тому +1

    This was very helpful. Thank you.

    • @joshemman
      @joshemman  2 роки тому +1

      Glad it was helpful!

  • @mohameddesouky4898
    @mohameddesouky4898 6 років тому +2

    Well Explained, if i have the Data and i should put the weight values how i can calculate them !
    i just have Data for sales last year and want to use this method it's applicable and how !
    thank you very much for your support

  • @mishellwadzanai92
    @mishellwadzanai92 Рік тому

    thank you so much this was really helpful, it was loud and clear...

  • @susah135
    @susah135 7 років тому +2

    Can we using the method for seasonal data where the irregularity can be extreme? For example, can we know that next week the sales would be 60, which is very different than the rest of your data?

  • @rickyc1578
    @rickyc1578 4 роки тому

    Correct me if I am wrong but for Forecast 5, the Numerator should read 3(45) + 2 (40), not 3 (45) + 2 (44) since we are using the last two most recent Sales (45 and 40). The answer 43 is correct though. Thanks by the way, you explain it so clearly!

    • @joshemman
      @joshemman  4 роки тому +1

      You're right. F5 should be (3(45) + 2 (40))/5 at 3:43. Thanks for pointing it out.

    • @rickyc1578
      @rickyc1578 4 роки тому

      @@joshemman normally if the instruction does not give a weighted value and youre asked to find the WMA for 2 periods, is it assumed its weighted by 1...so splitting the two periods by .6 and .4?

    • @joshemman
      @joshemman  4 роки тому

      @@rickyc1578 You should be given at least one weight.

  • @tanaman7111
    @tanaman7111 4 роки тому

    May the force be with you Joshua

  • @ellisakerragasajo7889
    @ellisakerragasajo7889 2 роки тому +1

    Thank you for this, Sir!❤️

  • @NaveenBali4
    @NaveenBali4 6 років тому +2

    Thanks for this simple and straightforward explanation.
    I have a question on when do we use MSE vs MAE for error calculation

  • @Brownskin-x5t
    @Brownskin-x5t 2 роки тому +2

    You saved me.
    Merci beaucoup!

  • @jonathancaicedo5105
    @jonathancaicedo5105 5 років тому +2

    Your videos are amazing and really easy to understand, do you have anything about Simple Linear Regression?

  • @bibekanandabhattacharjee952
    @bibekanandabhattacharjee952 3 роки тому

    That was very comprehensible. Like it.

  • @mussamageke
    @mussamageke Рік тому +1

    Very good explanation

  • @nathaliyaminukshi8630
    @nathaliyaminukshi8630 4 роки тому

    thank you for the explanation. this was soo easy to understand.

  • @yearsemakid2569
    @yearsemakid2569 4 місяці тому

    Mister my question the instruction only weight moving average approch by your given example how to calculate?

  • @anuragsrivastav2131
    @anuragsrivastav2131 6 років тому +2

    Hello Joshua,I would like to thank you for wonderful explanation.i have a question will the smaller value of m will give good forecast if so then how.

  • @eriol33
    @eriol33 3 роки тому

    if I got my next job, I must thank you

  • @averyperosa3052
    @averyperosa3052 4 роки тому +1

    thank you very much for this video, why is 45 the most recent value to start?

    • @joshemman
      @joshemman  4 роки тому +1

      Of all the recent periods, 45 is the closest the period we want to forecast.

  • @RaynerGS
    @RaynerGS 5 років тому

    #Weighted Moving Average 4WMA
    weight = [0.4, 0.3, 0.2, 0.1]
    sales = [39, 44, 40, 45, 38, 43, 39]
    calc = lambda x: 0 if x < len(weight) else sales[x-1]*weight[0] + sales[x-2]*weight[1] + sales[x-3]*weight[2] + sales[x-4]*weight[3]
    forecast = [ calc(x) for x in range(0, len(sales) ) ]
    print(forecast)

  • @freddy1599
    @freddy1599 4 роки тому

    Thank you. Which formula can I use in predicting the overhead budget.
    Thanks again

  • @novagamer6198
    @novagamer6198 3 роки тому

    Thank you so much I understand everything keep up the good work 👍

  • @emanfahmi84
    @emanfahmi84 5 років тому

    Big thanks you are GREAT , VOICE, VERBAL , SMOOTH EXPLANATION, THANK YOU :)

  • @hassanalianees5530
    @hassanalianees5530 3 роки тому

    What is the benefit of taking Moving Averages?
    How can we interpret MAD?

  • @pibidenalanka987
    @pibidenalanka987 Рік тому

    Thank you very much, Sir for explaining well.

  • @gooddaniel8603
    @gooddaniel8603 8 років тому

    Very incredible, wish you include more topic in that video lecture

  • @talatuhassan6469
    @talatuhassan6469 6 років тому +1

    How to find exp entail smoothing regression analysis, Markov chain, simulation. Thank you

  • @anchalmishra
    @anchalmishra 3 роки тому +1

    on what basis you calculated weights

  • @RichaHanda
    @RichaHanda 8 років тому

    if i increase number of weeks then error will decrease or increase. if error is increase as shown in ur example that 2wma give better smoothing than 4 wma then it means as number of periods increases the smoothing error get increased.

    • @joshemman
      @joshemman  8 років тому

      Not at all. The errors depend largely on the data you have.

  • @푼야와
    @푼야와 3 роки тому

    How about there is no given weight average but we need to calculate the three months and six montsh weighted moving averages?. Can someone help me/enlighten me please. THANK YOU

  • @AnAyahaDay
    @AnAyahaDay 4 роки тому

    What if we want to forecast more than one values in future. For example 21 future values in future? considering we have enough past data.

    • @joshemman
      @joshemman  4 роки тому

      You cannot forecast more than one period into the future with this method.

  • @Rambowayne
    @Rambowayne 7 років тому +1

    Man, good job. It was helpful.

  • @hassanalianees5530
    @hassanalianees5530 3 роки тому

    How can we know what weights should be used?

  • @sushmabalakrishnan4909
    @sushmabalakrishnan4909 7 років тому +1

    Hello.. The video is great explained very clearly
    But I had a doubt, when "we assume the weights of different values the sum of their weights should be equal to 1", Here It is satisfied for 4week but not in case of 2week??

  • @ntakirutimanaeric5819
    @ntakirutimanaeric5819 11 місяців тому +1

    thank you for your deeply help.

  • @AdamELclon
    @AdamELclon 6 років тому

    well explained and very helpful, Thank you so much Joshua

  • @nikkinikki8745
    @nikkinikki8745 5 років тому

    Please Josh how would interpret the 2.92 result?

  • @sophietan2652
    @sophietan2652 4 роки тому

    Great explanation!! Thank you so much :)

  • @mbekezelencube437
    @mbekezelencube437 3 роки тому

    HOW DO I CALCULATE THE FORECAST GIVEN 3 WEIGHTS ONLY?

  • @ThePhysics1234
    @ThePhysics1234 6 років тому

    How do you differentiate Mean absolute deviation vs Median absolute deviation which is also short for MAD

  • @Firestarter009
    @Firestarter009 7 років тому

    hi joshua
    could you further evaluate for Normalised weighted Root Mean Squared logarithmatic Error,please

  • @hSep-or5sh
    @hSep-or5sh 7 років тому +1

    thank you for this amazing explanations 👌

  • @HealthyFoodBae_
    @HealthyFoodBae_ 3 роки тому

    How do you know which weight values to use?

  • @linhnguyen-qr6dp
    @linhnguyen-qr6dp 3 роки тому

    Thank you so much for the videos. But I’m wondering if we could calculate week 8-10? how could I do that?

    • @joshemman
      @joshemman  3 роки тому +1

      No, you cannot. You can only forecast one period into the future using this method.

  • @alviandarif3220
    @alviandarif3220 4 роки тому

    Thank you, that was useful for my study!

  • @nourkhawari5168
    @nourkhawari5168 5 років тому

    what if the ( 4 weeks ) is not given, the question only says find the forecast for a specific month? how can i solve ?

  • @OmarAl93
    @OmarAl93 2 роки тому

    I have a small question on which basis should we consider the weights e.g. 2WMA we have picked 2,3. In 4WMA we have chosen 0,4 0,3 0,2 0,1 relatively. Is there a reseaon for choosing the weights? Can the column 4WMA values represent the upcoming year forecast?

    • @joshemman
      @joshemman  2 роки тому +1

      The reason for choosing the weights can depend on the nature of the data, the objective of the forecast, and relevance of the recent data. But it is nothing to worry about unless you’re trying to choose your own weights.
      Yes! The 4WMA values are the forecast values.

    • @OmarAl93
      @OmarAl93 2 роки тому

      @@joshemman Thanks a lot, this makes a lot of sense now :). I applied the method on 2021 to forecast 2022, the results are below the actuals I have in 2022 (Jan-Sep) Would you recommend more statistical methods to have more accurate results? I already have data for 2021 and 2022.

    • @joshemman
      @joshemman  2 роки тому +1

      ​@@OmarAl93It depends on your data and your objective, Omar. Weighted moving average is usually applied to data that has no trend (not consistently increasing or decreasing). If there is trend, then regression might be a better approach.

  • @chrislainma5997
    @chrislainma5997 5 років тому

    Very good explanation, thanks

  • @mayceperalta520
    @mayceperalta520 Рік тому

    Thank you for thisssss 🥰🥰

  • @kingngote5201
    @kingngote5201 5 років тому

    How can you calculate the sales for week 9 and 10.

  • @MindBeliever
    @MindBeliever 4 роки тому

    Anyone know how it work on "optimum weighted moving avarage". There is no weights?? 🤔

  • @froilansalac6030
    @froilansalac6030 6 років тому

    thanks! i Got great ideas for this videos!

  • @ArslanJapani
    @ArslanJapani 4 роки тому

    That Really helped alot sir!
    Thankew

  • @FuatEnesARICI
    @FuatEnesARICI 5 років тому

    how to decide of their weights ??

  • @Eussub40
    @Eussub40 2 роки тому

    What is you were asked in a question like what would be the weighted moving average for week 6 if the weight were 0.40, 0.30, 0.20 and 0.10?

    • @joshemman
      @joshemman  2 роки тому

      You assign the weights respectively to periods 5, 4, 3 and 2.

    • @Eussub40
      @Eussub40 2 роки тому

      ohhh so then the weights would be 0.5, 0.4,0.3,0.2 and 0.1?

    • @joshemman
      @joshemman  2 роки тому

      @@Eussub40
      0.40, 0.30, 0.20 and 0.10.
      Where did 0.5 come from?

    • @Eussub40
      @Eussub40 2 роки тому

      @@joshemman I really don't know how to work out this equation. I'm doing an entrepreneurship course and I hate this part. The reason I said 0.50 cause you said assign the weight respectively to period 5 etc

    • @Eussub40
      @Eussub40 2 роки тому

      @@joshemman maybe I misunderstood

  • @kermit8556
    @kermit8556 2 роки тому +1

    you are AMAZING

  • @eeshahahsan3997
    @eeshahahsan3997 5 років тому

    What if the sum of weights is less than 1?

  • @jjjahed7272
    @jjjahed7272 8 років тому

    If there three weights in question but two of them are given.
    How can I calculate the third weight, since it wasn't given. Thanks

    • @joshemman
      @joshemman  8 років тому

      +MD Jahedul Islam
      Just make sure the they add up to 1 (or 100%).

  • @Deebz786
    @Deebz786 9 років тому

    Joshua you are the man

  • @melprincorrea4237
    @melprincorrea4237 4 роки тому

    how do you come up with weights of .4.3.2.1 for 4week Weighted average and for 2 week weighted average , how did u get Weights 3 and 2? is it an automatic assumption?

    • @joshemman
      @joshemman  4 роки тому

      Yes. They are assumed (chosen randomly for the example)

  • @TrangNguyen-si8ru
    @TrangNguyen-si8ru 4 роки тому

    Hi, How can we decide the weights?

  • @axineriego1674
    @axineriego1674 2 роки тому

    If the given weight equals to 1 means you don't have to dividw it by its n, then if the weight is not equal to 1 then you have to divide it by its n, right?

    • @joshemman
      @joshemman  2 роки тому +1

      Yes. If the weights add up to 1, you're essentially dividing by 1 which appears like doing nothing.

    • @axineriego1674
      @axineriego1674 2 роки тому

      @@joshemman Thank you.

  • @dddaaev
    @dddaaev 5 років тому

    Hello! I got a question, I was reading about this method, and I saw that all the weights must be equal to 1, I'm a bit confused why did you choose weight 5 on the second example.

    • @ahboiiyan5337
      @ahboiiyan5337 5 років тому

      its just an example i think is it does not equal to 1

    • @meuko
      @meuko 5 років тому

      It's true that the total weight must be equal to 1, but he explains that even though his weight totals at 5, he divides his final answer by 5, so he gets the actual weight of one.

  • @ratupangpang7329
    @ratupangpang7329 3 роки тому

    Why use weights 3 and 2 at 2WMA instead of using weights 2 and 1?

    • @joshemman
      @joshemman  3 роки тому

      The weights are assigned based on the level of importance recent periods. They will vary from one problem to another.

  • @haticekubragurses8568
    @haticekubragurses8568 6 років тому

    can u explain decomposition on your nex video

  • @yourangel7038
    @yourangel7038 4 роки тому

    what if weight was not given ? what will i assume that the weight will be?

    • @joshemman
      @joshemman  4 роки тому

      If the weights are not given, then it's just simple moving average. The weights will be 1 for each period.

  • @andrewlian9866
    @andrewlian9866 8 років тому

    In 2-week WMA case, weight 3 and 2, are they just random number? and how to get the optimal weights? thanks

    • @joshemman
      @joshemman  8 років тому

      +Andrew Lian
      You can use Excel's Solver to obtain optimal weights. See ua-cam.com/video/zOGfG65ASdA/v-deo.html

    • @andrewlian9866
      @andrewlian9866 8 років тому

      Joshua Emmanuel thank you! that would be helpful

  • @joelopez3552
    @joelopez3552 4 роки тому

    i love this dude

  • @axineriego1674
    @axineriego1674 2 роки тому

    Sir I have another question, the lower the MAD and MSE, the forecasting method used is more accurate?

  • @monikakamra679
    @monikakamra679 8 років тому

    kindly solve this Ques ..it's taken from one of the prestigious competitive exam of India..so the Ques Is- given the numbers 2,6,1,5,3,7,2. if the weights used are 1,4,1 .then weighted moving average of order 3 are given by?.

    • @joshemman
      @joshemman  8 років тому

      2,6,1,5,3,7,2 with weights 1, 4, 1. The sum of the wights is 6. For weighted moving average of order 3,
      F4=(2+4*6+1)/6 = 4.5
      F5=(6+4*1+5)/6 = 2.5
      ...
      F8=(3+4*7+2)/6 = 5.5

    • @monikakamra679
      @monikakamra679 8 років тому

      Can u explain the same in ur video.?.I would be very grateful to u.. thank u for teaching it beautifully..

  • @Dan71Pel
    @Dan71Pel 6 років тому

    Great vid. Thank you so much!

  • @jeffjericcastillo3710
    @jeffjericcastillo3710 3 роки тому

    hi sir, may i ask what weights to use in 3-WMA?

    • @muathsaeed1569
      @muathsaeed1569 3 роки тому

      the weights are decided by the management who conducted the study based on what they see is important. they are not fixed numbers, you could choose any weights you want. but generally managemnet or the person who is doing the forecasting give the recent data more priority than old data which means higher weights for rececnt data

  • @hydershaikh83
    @hydershaikh83 8 років тому

    so finally there are 2 types of method if we got value 1 by calculating all given weight value so we have to have solve from 1st method if we are getting another value like 5 or 7 or 10 by calcuting so we have to do by the second method

  • @nicktelsa3473
    @nicktelsa3473 6 років тому

    So what happens to the first 4 weeks that are empty? Where do we get that data from?

    • @joshemman
      @joshemman  6 років тому

      Nothing happens. You can't get data for those periods so you leave them alone.

  • @yearsemakid2569
    @yearsemakid2569 3 місяці тому

    Mister i have a question F5 how to calculate i think writing error 2WMA
    F5=45*3+40*2/5 but u write F5=45*3+44*2/5➡️43 Edit ur video please

    • @joshemman
      @joshemman  3 місяці тому

      Sorry about the typo. Unfortunately, the video cannot be edited. But I think you get the idea.