MATH38172 Generalised Linear Models
Semester 2, 2023/24
Coursework
Background: The ile drill.csv on Blackboard contains data originally described by Cuth- bert Daniel in his 1976 book on industrial experimentation. It contains data on response Y - the advance rate of the drill and four factors, X1 = Drill load, X2 = Flow of a lubricant, X3 = Speed of drill, X4 = Type of mud. Each of the four factor variables are recoded at two levels each, “High“ and “Low“ .
(a) Describe a generalized linear model to the advance rate response with gamma as response distribution, include all four regressors in the linear predictor and use the canonical link function. [3]
(b) Use R to it the model in (a), and interpret the signs of the parameter estimates. [3]
(c) Fit a second gamma glm for the same response Y, with all four covariates, using the log-link- function and interpret the signs of the parameter estimates. [3]
(d) Compare the two models itted in parts (b) and (c). Which of the two models you recommend and why. [2]
(e) Briely explain two alternative ways of testing signiicance of a parameter and apply them to the model itted in (b) for the parameter corresponding to the dummy variable for the covariate X1 . [4]
(f) Explain which of the two methods from (e) might be better. [1]
(g) For the model itted in part (c) expand the linear predictor to include all six of the two - factor interactions involving the four covariates. Compare the model deviance for this model to the model deviance from part (c). Are the additional terms in the model useful in terms of model it? Justify your answer. [4]
[Total 20 marks]
版权所有:编程辅导网 2021 All Rights Reserved 联系方式:QQ:99515681 微信:codinghelp 电子信箱:99515681@qq.com
免责声明:本站部分内容从网络整理而来,只供参考!如有版权问题可联系本站删除。