Dear Sir
I would like to know if anyone can help to apply Four-Parameter Logistic (4PL) and Five-Parameter Logistic (5PL) in excel
Dear Sir
I would like to know if anyone can help to apply Four-Parameter Logistic (4PL) and Five-Parameter Logistic (5PL) in excel
more details about equation from this site
http://help.idbssupport.com/idbs/ser...26curagt%3Dndx
Please post an Excel file (not image).
this is the data file
I am no mathematician but a very quick read of 4PL suggests you need 4 parameters so as you supplied 2 "parameters" it suggests to me (in my ignorance) you cannot use this technique.
There are respondents to this forum who are mathematicians so hopefully they will look at your request and offer advice.
Your file consists of 6 data points and nothing more. Here's how I would do this:
1) Enter my "initial guesses" for the parameters of the desired form of the logistics function. The 4PL is the same as the 5PL with g=1, so I'd probably set up the 5PL and decide later whether to fix g at 1 or let it float.
2) Enter the logistics function in an adjacent column. With the right combination of relative and absolute references, you should only need to enter the formula once, then copy it down.
3) I almost always include a column that computes some "deviation score" for each point. An absolute difference (yestimated-ymeasured) or a relative difference (yestimated/ymeasured-1) or something.
4) Choose an objective function (OF)-- the function you want the regression to maximize/minimize. Then enter that function into the spreadsheet. This might be the sum or product of the deviation score column or some other function. If you want the common "least squares" OF, you can use the =SUMXMY2() function: https://support.office.com/en-us/art...4-e2ecee23b299
5) Make sure Solver is installed, then call Solver and tell it to:
5a) Set target cell -- the cell with your OF
5b) To a minimum or maximum (whichever choice is appropriate for your chosen OF)
5c) By changing -- the cells where you entered the initial guesses for the parameters. Here is where you will choose whether to include g or not.
5d) Solve.
If you have given it reasonable first guesses, it should converge on a "best fit" of the data according to the OF you programmed in. What part of that do you get stuck on?
Originally Posted by shg
Dear Mrshorty
thanks for your concern
but i didn't understand because im not good in statistic.
i just want to make my own file for fitting curve like with all methods of fitting like https://mycurvefit.com/
i already get from internet file with two methods , i didnt know if low of forum allow to upload the file or not ?
I don't know why this forum would disallow uploading those files -- unless there is some kind of copyright or license on those files that prohibits uploading to the forum or something like that.
You may not be good in statistics, yet, but this is how you get better. I find that a lot of programming problems need to be approached by breaking things up into smaller steps, which is what I did above. The idea is to break it up into parts that you know how to do so we can focus on the parts you need to learn. I assume almost everyone knows how to do 1 -- enter values into cells. What part or parts do you not know how to do? Do you have trouble entering the logistics function for one point (step 2)? Or are you having trouble using Solver?
this is the model file i found on internet
Use Solver to minimize the RMS error by changing the five variables:
A B C D 1 a 2.420179 2 b 1.044576 3 cc 6.732798 4 d 0 5 g 0.61649 6 RMS Err 0.036506C6: {=SQRT(AVERAGE((C9:C14-B9:B14)^2))} 7 8 [std] Reading 9 0 2.417 2.420C9: =d + (a - d) / (1 + (A9 / cc) ^ b) ^ g 10 4 1.852 1.825 11 10 1.312 1.372 12 25 0.955 0.904 13 60 0.569 0.557 14 120 0.336 0.368
Entia non sunt multiplicanda sine necessitate
Dear Shg
how can i applied this in my sheet( from where you get value for a,b cc,d,g
Data > Solver
http://www.solver.com/solver-tutorial-using-solver
There are currently 1 users browsing this thread. (0 members and 1 guests)
Bookmarks