Description
- Use the data set you used in HW 3 and answer the following questions.
- Do a LOESS fit for Calories using the variables: weight, height, protein, carbohydrates, calcium and gender and look at the relationship that LOWESS provides and comment.
- Take the residuals versus predicted plot that you ran for HW3 diagnostics and LOWESS it. Do you see any patterns? Should you?
- You are given the data on time of exposure(x) and photolytic damage (y) for an exposed surface on the web as DATA for HW5_nlindata.
The scientists believe that the model relating damage and time is given by:
damage = Theta1 – Theta2 * exp(-Theta3 * time)
- Estimate the parameters theta1, theta2 and theta3 and test for the significances of the parameters. Use your estimates to predict y given x. What can you say about the model?
- Use bootstrapping to verify your results.
- Do a LOWESS plot to compare your linear and non-linear fit.