Please use:

install.packages("MASS")

library(MASS)

data(ships)

to get access data in r studio.

Three problems can be seen as one.

1. The ships dataset in the MASS package gives the number of wave damage incidents and aggregate months of service for different types of ships broken down by year of construction and period of operation. a. Fit a Poisson regression model for the number of incidents with predictors: log of service, type, year and period. Treat year and period as categorical predictors. Find the p-value for the Wald test that the coefficient for log of service years is 1. What is the interpretaion when this parameter is 1? b. Fit a Poisson rate model with an appropriate offset with predictors type, year and period. Find a 95 percent confidence interval for the parameter e where B is the coefficient for the dummy variable for being ship type B. c. Check the standardized Pearson residuals. Are there any outliers? 2. a. Estimate the overdispersion parameter for the model in (1b). b. Fit a quasi-Poisson model with the same offset and predictors as (1b) allowing for overdispersion. Once again compute the confidence interval for es of (1b). How has it changed? 3. Interpret the coefficients of the quasi-Poisson model. What factors are associated with higher and lower rates of damage incidents