Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.
Showing results for
Search instead for
Did you mean:
Our World Statistics Day conversations have been a great reminder of how much statistics can inform our lives. Do you have an example of how statistics has made a difference in your life? Share your story with the Community!
I am currently doing a analysis on oven profile. We have 16 data points from the oven from thermocopules. and 2 maximum data repreenting the Max Hot Temperature and Max cold temperature .
the 16 data points and 2 Maximum data points are output responses. Which means there is two sets of output response - 1) Profile temperature from thermocouple 2) Oven display temperature from 2 sensors.
Which analysis is best for predicting the actual profile temperature using these two data sets?
Correlation or Regression Fit model? Since both of it is a output response i am confused on this.
Since this is a curve response, i find that using quadratic regression model is a better fit. Any help on the correct analysis method for this case is useful.
Yes, 'input' or 'output' is all a matter of one's point of view!
I have a general idea from your description, but I think you need to be more specific when you say '... analysis on an oven profile'. What is the practical objective of such an analysis? To take an example, it's not uncommon to want to control a temperature profile, or maybe even make it more uniform, because that affects what comes out of the oven. In the case of 'control', that may or may be possible using just the recorded max and min values (and not the 16 others). Whether this possibility is reasonable could be determined by an analysis that seeks to understand the relationship between the 16 measurements and the 2 others. In that analysis the 16 would be thought of as inputs, and the 2 as outputs.
Once this is clear, folks should be able to offer some suggestions about what techniques might be good for determining the relationship of interest. As usual, sample data in a JMP table would help, if you are able to share some.
Attached is the Oven thermocouple readings and temperature readings from the oven display.
Basically in our study we are actually trying to see the relationship between the actual thermocouple readinsg we get from the oven chamber vs the two readings which is displayed on the oven. We have recorded the max values of our thermocouple readings and the oven display readings. now we are trying to do is to predict the oven chamber readings with only the display readings. We are collecting more data. However my question if a quadratic fit regression model will be the appropriate one. Will they be considered a dependent variable? is the regression model fit the correct statistic method for this?
Your .pptx was quite informative, but just contained a screenshot of the measured values so I could only get a general impression. Clearly, more data would be good, and TBH, I don't think one could make reasonable recommendations until such data is available.
In cases like this, and before thinking about modeling, it's essential to look at the data graphically to understand better the sources of variation it contains. The oven runs you do currently have came from several different conditions of 'Tray', 'Type' and 'Sequence', and (if I take the colour coding seriosly and ignore any questions about measurement uncertainty), it looks like the pattern of termperatures within the oven may depend on these conditions. So you will need to decide if your 'prediction' (using a statistical approcah not yet specified) is intended to cover all cases together, or only some. In any case, no matter what technique you finally use, getting sixteen numbers from two seems ambitious.