I'm looking to develop a master curve considering three parameters, any suggestions for statistical model, like Sigmoidal function, any other such model which will be helpful?
dear Dr. Gill,
basically i have different curves with same slope only differing the intercept. and i also figured out that this intercept change is because of another parameter. so, i'm thinking to develop a single master curve which can represent all the conditions.
Dear Dr. Gill.
I have attached my data set,in future i will have more similar data set. Please suggest me the way i'm thinking to represent these different curves by single master curve will work or not. It's actually about developing a new friction design guidelines in road construction. I highly appreciate your concern about my questions.
Ah ha, this looks like a regression problem. You want to predict y="DFT20" (numeric) from x="1000s of polish cycles" (numeric) and "type?" (categorical, three categories). You only have observations at five different values of x. One sees per category that y tends to decrease with x though not linearly, it looks as though there is a horizontal asymptote. And it looks as though the three curves are (vertical) translations of one another.
So why not fit some kind of linear or non-linear regression model to all the data? Assuming that the effects of "polish cycles" and "type" are additive?
Yes, but at the end of the day what are you going to compute: maybe the inflection point? (if it happens so?)
Dear Dr. Gill,
Thank you for the suggestion, i will work on my more data may be after this week and try to develop the model as you suggested and will again come to discuss. Any suggestion of non-linear model for this kind of data pattern.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics