You can access the horseshoe crab data by installing in R with this code install

Question

You can access the horseshoe crab data by installing in R with this code

install.packages("icda",repos="http://www.stat.ufl.edu/~presnell/R",type="source")

install.packages(c("reshape2","VGAM","gee"))


library(icda)


data(horseshoecrabs)





Refer to the horseshoe crab data.
a. Report the fit for the probit model, with weight predictor.
b. Find Pi_hat at the highest observed weight, 5.20 kg.
c. Describe the weight effect by finding the difference between the Pi_hat values
at the upper and lower quartiles of weight, 2.85 and 2.00 kg.
d. Interpret the parameter estimates using characteristics of the normal cdf
that describes the response curve.

Thanks a lot!

Explanation / Answer

For clarification of this question read question no 3.7 and you will get to know that we need to make a probit model for number of satellites where Y=0 for satellite=0 Y=1 for satellite>=1

install.packages("icda",repos="http://www.stat.ufl.edu/~presnell/R",type="source")
install.packages(c("reshape2","VGAM","gee")) library(icda) data=horseshoecrabs

library(MASS) p=data$Satellites p[which(p>=1)]=1 myprobit <- glm(factor(p)~ data$Weight, family = binomial(link = "probit")) summary(myprobit)
a). pi_hat = -2.238 + 1.099*(Weight)

b). myprobit$fitted.values[which(data$Weight==5.2)] Ans = 0.9997462

c). myprobit$fitted.values[which(data$Weight==2.85)] = 0.8143302
myprobit$fitted.values[which(data$Weight==2.0)] = 0.4839603
difference=0.8143302-0.4839603 = 0.3303699


d).we need to fir normal distribution over data obtained for pi_hat
for that we need to install a package
install.packages("fitdistrplus") norm=fitdist(myprobit$fitted.values, "norm")
Parameters: estimate Std. Error mean 0.6420447 0.01477920 sd 0.1943901 0.01044923 Parameters: estimate Std. Error mean 0.6420447 0.01477920 sd 0.1943901 0.01044923
so pi_hat = CDF of normal distribution for given mu and sigma.

Related Questions

Navigate

• Browse (All)
• Browse Y
• Subjects

---

---

Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.