Bird Species in Solomon Islands A study looked at the relationship between: ο· S = # of Species living near sea-level on the individual islands of Solomon archipelago ο· A = Area in miles2 ο· D = Distance from center of archipelago in miles (many have 0 for value) The authors fit the model: −π·π½2 } π½3 π΄π½4 Fit the regression model. As a starting set of values, setο ο’ο°ο½ο³ο΅ο¬ο ο’ο±ο½ο΅ο¬ο all other ο’s = 1.This is based on fitting a simple linear regression between S and log(A) for the islands with D=0, to get starting values for ο’ο°ο and ο’ο±. πΈ(π) = (π½0 + π½1 log(π΄))ππ₯π { Formula: species ~ (b0 + b1 * logarea) * exp(-(dist^b2)/(b3 * area^b4)) Parameters: Estimate Std. Error t value Pr(>|t|) b0 34.85994 0.56682 61.501 < 2e-16 b1 5.34139 0.12131 44.031 < 2e-16 b2 0.66986 0.07581 8.836 1.49e-11 b3 11.30165 3.26559 3.461 0.00116 b4 0.28720 0.03357 8.556 3.82e-11 *** *** *** ** *** Residual standard error: 3.275 on 47 degrees of freedom Complete the fitted values for the following combination of values of A and D: A\D 0.5 2 10 1000 0 38.56 47.16 71.75 25 12.23 20.63 31.86 64.71 75 4.42 10.45 20.79 57.85 225 0.53 2.52 8.53 sol <- read.fwf("http://www.stat.ufl.edu/~winner/data/solomon.dat", width=c(20,3,8,8), col.names=c("island", "species", "area", "dist")) attach(sol) logarea <- log(area) dist <- dist+0.001 species.mod <- nls(species ~ (b0+b1*log(area))*exp(-(dist^b2)/(b3*area^b4)), start=c(b0=35,b1=5,b2=1,b3=1,b4=1)) summary(species.mod) dist.grid <- 0:350 plot(species ~ dist) lines(dist.grid,predict(species.mod,list(area=0.5,dist=dist.grid)),lty=1) lines(dist.grid,predict(species.mod,list(area=2,dist=dist.grid)),lty=2) lines(dist.grid,predict(species.mod,list(area=10,dist=dist.grid)),lty=3) lines(dist.grid,predict(species.mod,list(area=1000,dist=dist.grid)),lty=4)