Fit Nonlinear Model to Data with Excel

แชร์
ฝัง
  • เผยแพร่เมื่อ 1 ธ.ค. 2024

ความคิดเห็น • 53

  • @mjmeternal2696
    @mjmeternal2696 ปีที่แล้ว +1

    I learnt a lot especially that point where you imported data.

  • @SH-pc4xt
    @SH-pc4xt 4 ปีที่แล้ว +1

    Thanks very much for posting this helpful into tutorial to using Solver in Excel. One thing that seemed odd was referring to the equation you're fitting as 'the correlation'. It's not a correlation, it's an equation with parameters. When these are suitably adjusted the result may (or may not) produce a calculated result that matches the original data.

    • @apm
      @apm  4 ปีที่แล้ว +2

      Thanks for that clarification. Exact wording is important to distinguish between a statistical definition of "the process of establishing a relationship or connection between two or more measures" and a more common definition of "a mutual relationship or connection between two or more things". It is common in some disciplines to refer to equations with parameters adjusted to minimize a measure of closeness as correlations. I like your definition because it is more specific and doesn't create any ambiguity across disciplines.

  • @niguvoice2511
    @niguvoice2511 ปีที่แล้ว +1

    If I want to do fittings on a 2Y graph on excel. What should I do?

    • @apm
      @apm  ปีที่แล้ว

      For multiple regression try Python: apmonitor.com/pds/index.php/Main/RegressionOverview

  • @amirashkansaadati4480
    @amirashkansaadati4480 3 ปีที่แล้ว +1

    Thank you so much, would you please talk about the best way to find right formula for some nonlinear data? I mean when we have data, how we will be able to find their relationships based on formula?

    • @apm
      @apm  3 ปีที่แล้ว

      If you don't have a particular physics-based form then you could try many different forms and see which one does the best at fitting the function. Otherwise you could use machine learning and let the computer decide the form. apmonitor.github.io/data_science

  • @murrowboy
    @murrowboy 3 ปีที่แล้ว +2

    Dumb question but how accurate is regression compared to just taking a moving average of previous values?

    • @apm
      @apm  3 ปีที่แล้ว

      Regression gives you a model that you can use to predict future changes. Moving average could also be a future predictor but it suggests that future values are similar to recent values.

    • @murrowboy
      @murrowboy 3 ปีที่แล้ว

      @@apm I'm assuming this is a popular technique for stock prediction. I wonder how accurate it can be.

  • @anasforum7101
    @anasforum7101 ปีที่แล้ว +1

    Thank you very much, it's a fantastic tutorial, I love it.

  • @alvafla
    @alvafla 5 ปีที่แล้ว +1

    Very useful video, I did it with a "complex" curve and fitted it very well (impossible with polynomial equation)... thank you

  • @CodyVLopez
    @CodyVLopez 4 ปีที่แล้ว +2

    Wow this is awesome, thank you for making this! Didn't know that solver could make it so easy

    • @apm
      @apm  4 ปีที่แล้ว

      It is a great tool for small to medium sized problems in optimization.

  • @blrd3277
    @blrd3277 4 ปีที่แล้ว +1

    Great video. Need more info for excel after this step is done.
    Could you suggest how to estimate quality of the fit by not only adding Pearson's coefficient, but also standard errors for each variables, c1, c2, c3, c4. This essentialy equalize Excel to other fitting software like Kaleidagraph and Origin. If we can draw various confidence bounds in Excel and have also "third type of axis" -Normal (Excel only ahve linear and log axis)- then we can approach possibiliies of Minitab. Thanks. Being able to do eveything from "scratch", rather then rely on software calcs is invaluable

    • @apm
      @apm  4 ปีที่แล้ว

      Some of those analyses are only valid for linear relationships: stats.stackexchange.com/questions/220080/pearsons-correlation-for-non-linear-data There is a nice regression add-on to Excel called "Data Analysis" that does most of the basic analysis types for linear regression. Here is more advanced material for nonlinear regression (in Python): apmonitor.com/che263/index.php/Main/PythonRegressionStatistics

  • @robinoconnor5760
    @robinoconnor5760 4 ปีที่แล้ว +1

    This was extremely helpful. Is there a similar method for multivariable regression?

    • @apm
      @apm  4 ปีที่แล้ว

      Yes, but you may want to use a tool like Matlab or python apmonitor.com/me575/index.php/Main/NonlinearRegression

    • @apm
      @apm  4 ปีที่แล้ว

      Here is more content apmonitor.com/me575/index.php/Main/LinearMultivariateRegression

  • @oxydol3456
    @oxydol3456 3 ปีที่แล้ว +1

    Great video. Do you think it's not easy to create the solver-function clone from scratch?

    • @apm
      @apm  3 ปีที่แล้ว

      A basic solver would be possible. Here are some guiding principles: apmonitor.com/me575

  • @KamalSingh-dn7gv
    @KamalSingh-dn7gv 3 ปีที่แล้ว

    Excellent video. Thank you.

  • @suryavamsi9201
    @suryavamsi9201 6 ปีที่แล้ว +4

    Thank you so much, well explained

  • @mahmoud_ali_963
    @mahmoud_ali_963 2 ปีที่แล้ว

    thank you man you are the best :D

  • @yuziwang7705
    @yuziwang7705 5 ปีที่แล้ว +1

    Hi, thank you, it's very helpful. So we can actually use the RSQ function to check how good the fitting is even it's a nonlinear model, right? Because I was not sure if the RSQ could be only used for a linear model or not.

    • @apm
      @apm  5 ปีที่แล้ว

      Yes, RSQ works for linear or nonlinear regression.

  • @fowul
    @fowul 4 ปีที่แล้ว

    Can this applies to multi variable non linear regression as well?

    • @apm
      @apm  4 ปีที่แล้ว

      Yes, you can do multivariate nonlinear regression as well. The Excel solver won't do very well with big data or very nonlinear equations. You may want to try a solver package like Gekko in Python: apmonitor.com/me575/index.php/Main/NonlinearRegression

  • @mohammadalikeshtiara3573
    @mohammadalikeshtiara3573 4 ปีที่แล้ว +1

    Thank you so much. really helpful

  • @anisizyanilee5035
    @anisizyanilee5035 5 ปีที่แล้ว +2

    Thank you so much! Very helpful.. God bless you :)

  • @miguelangelsanchez5959
    @miguelangelsanchez5959 6 ปีที่แล้ว

    7:34 Very good explanation, thank you.

  • @nicolasmoreno4990
    @nicolasmoreno4990 6 ปีที่แล้ว

    Do you know how to calculate the error of the parameters with excel?

    • @apm
      @apm  6 ปีที่แล้ว

      Two popular ways are sum of squared error or sum of absolute error.

  • @DigitalicaEG
    @DigitalicaEG 6 ปีที่แล้ว +2

    Very helpful thank you

  • @eyberantezana8843
    @eyberantezana8843 4 ปีที่แล้ว +1

    Excelente, gracias!

  • @NalinJayaratne
    @NalinJayaratne 6 ปีที่แล้ว +1

    Well explained.

    • @apm
      @apm  6 ปีที่แล้ว +1

      Thanks for the encouraging comment, Nalin.

  • @ajay4forest
    @ajay4forest 4 ปีที่แล้ว

    How to fit two way regression ?

    • @apm
      @apm  4 ปีที่แล้ว

      Here is an example with 3 predictor variables: apmonitor.com/me575/index.php/Main/NonlinearRegression You can do the same thing in Excel, if needed.

  • @gabrielrt
    @gabrielrt 4 ปีที่แล้ว +1

    Thank you so much!

  • @9986916092
    @9986916092 5 ปีที่แล้ว +1

    THANK YOU VERY MUCH

  • @samriddhatuladhar3913
    @samriddhatuladhar3913 4 ปีที่แล้ว

    i can not thank you enough.

    • @apm
      @apm  4 ปีที่แล้ว

      I'm glad it helped.

    • @samriddhatuladhar3913
      @samriddhatuladhar3913 4 ปีที่แล้ว +1

      @@apm What if i dont have the experimental values , how can i predict the values only with the equations ?

    • @apm
      @apm  4 ปีที่แล้ว

      @@samriddhatuladhar3913 it is no problem if you don't have experimental values. You can just put your equation into Excel to calculate the predicted outcome.

  • @ThinhTran-yu9sn
    @ThinhTran-yu9sn 6 ปีที่แล้ว

    Thank you very much

  • @waimyokhing
    @waimyokhing 6 ปีที่แล้ว +1

    please with matlab...

    • @apm
      @apm  6 ปีที่แล้ว

      I repeat most examples in Excel, MATLAB, and Python. You can see the MATLAB content here: apmonitor.com/che263/index.php/Main/MatlabDataRegression

    • @waimyokhing
      @waimyokhing 6 ปีที่แล้ว

      but in this example show the regression model. i want to know autoregression model.
      thanks...

    • @waimyokhing
      @waimyokhing 6 ปีที่แล้ว

      this is my nonlinear autoregressive model ...
      x(i+1) = a*x(i) + b*exp(c*x(i)) + epsilon(i);

    • @apm
      @apm  6 ปีที่แล้ว

      Here are examples of autoregressive modeling in Excel, MATLAB, and Python: apmonitor.com/do/index.php/Main/ModelIdentification I recommend that you look at the Excel example first because it is the easiest one to understand.

    • @waimyokhing
      @waimyokhing 6 ปีที่แล้ว

      thanks alot for your reply i try it....

  • @areejiman1304
    @areejiman1304 3 ปีที่แล้ว +1

    Thank you so much