I have read some paper on hydrology models which are calibrated using Response Surface method (RSM). After reading those papers, I have some question related to RSM. The model I am working takes input data as follows : Daily Precipitation data, daily temperature data and daily snow cover area. With this model, I have associated parameter/ factor ( 5 factors - Continuous) with the model. The accuracy of the model is determined by Nash-Sutcliffe coefficient (R square) using the computed and measured discharge from the model. I have to keep on changing those factors ( trail and error method ) in order to achieve high R sqaure value. (Dailymeasured discharge validates the accuracy by comparing with value from daily computed discharge generated by the model) Is it possible to find a optimum value for achieving target R square by this DOE technique? Please advice me.
Yes, the ability to generate a response surface design, and the ability to represent data with a response surface model. are native features of JMP; you don't need the PRO version for that.
Short answer: yes. Slightly longer answer: use the FIt Model platform to create models for each response; then use the prediction profiler to create a desirability function based on the goals for each response (maximise, minimise, achieve target). Then optimise the desirability function.
@@kirankumarkr4611 Yes, the ability to generate a response surface design, and the ability to represent data with a response surface model. are native features of JMP; you don't need the PRO version for that.
Well in the real-world you would conduct the experiment to collect the data, but for demonstration purposes I just used a formula to generate some simulated data.
For 2024 I am planning some new video content relating to DOE. What specific topics would be of interest? Please let me know if the comments :)
RSM CCD design and its analysis to predict the response.
You have a wonderful voice. Other than that, helpful video!
It's very useful video, thanks!
Does it work with canopy size as response variable and Altitudes gradient as explanatory variable?.. Thanks in advance
Please tell me how to determine and analyse the prediction equation in factorialdesign
Nice video, helped a lot..Thanx dear
How did you found Y values?
how do you simulate the data ? Ive been trying to figure that out for training/demonstration purposes
It was a column formula. Use the RandomNormal() function to add noise.
I have read some paper on hydrology models which are calibrated using Response Surface method (RSM).
After reading those papers, I have some question related to RSM. The model I am working takes input data as follows : Daily Precipitation data, daily temperature data and daily snow cover area. With this model, I have associated parameter/ factor ( 5 factors - Continuous) with the model.
The accuracy of the model is determined by Nash-Sutcliffe coefficient (R square) using the computed and measured discharge from the model.
I have to keep on changing those factors ( trail and error method ) in order to achieve high R sqaure value. (Dailymeasured discharge validates the accuracy by comparing with value from daily computed discharge generated by the model)
Is it possible to find a optimum value for achieving target R square by this DOE technique? Please advice me.
Does jmp and jmp Pro version have rsm tool?
Yes, the ability to generate a response surface design, and the ability to represent data with a response surface model. are native features of JMP; you don't need the PRO version for that.
Can jmp software used for multi response optimization
Short answer: yes. Slightly longer answer: use the FIt Model platform to create models for each response; then use the prediction profiler to create a desirability function based on the goals for each response (maximise, minimise, achieve target). Then optimise the desirability function.
Thank you sir, one more question, is the rsm tool present in both jmp and jmp Pro version
@@kirankumarkr4611 Yes, the ability to generate a response surface design, and the ability to represent data with a response surface model. are native features of JMP; you don't need the PRO version for that.
@@davidburnham3062 thank you
How you get simulation data when you make a table for y?
Well in the real-world you would conduct the experiment to collect the data, but for demonstration purposes I just used a formula to generate some simulated data.
How did you found Y values?