In deruncie/PLS205_package: Provides some functions to help with PLS205 assignments

library(ggplot2)
library(PLS205)

data = data.frame(G1 = rep(1:4,each = 20),G2 = factor(rep(1:2,each = 10)))
data$xx = runif(nrow(data)) + data$G1
data$y = .1*data$xx + (data$xx-mean(data$xx))^2 + .1*rnorm(nrow(data)) + data$G1 + as.numeric(data$G2)
data$G1 = factor(data$G1)
data$G2 = factor(data$G2)

p = ggplot(data,aes(x=xx,y=y)) + geom_point() + facet_wrap(~G1,scales = 'free')
p + geom_smooth(method = 'lm',aes(group = G2))

my_lm = lm(y~G1*xx*G2,data)
summary(my_lm)
pred_data = predict(my_lm,newdata = data,se.fit = T)
data$y_pred = pred_data$fit
data$y_max = data$y_pred + 2*pred_data$se.fit
data$y_min = data$y_pred - 2*pred_data$se.fit
p + geom_ribbon(data = data,aes(group = G2,ymin = y_min,ymax = y_max),alpha = .2) + geom_line(data = data,aes(color = G2,y = y_pred))

pred_data = predict(my_lm,newdata = data,se=T)
data2 = data
data2$y_pred = pred_data$fit
data2$ymax = data2$y_pred + 2*pred_data$se.fit
data2$ymin = data2$y_pred - 2*pred_data$se.fit

p %+% data2 + geom_ribbon(aes(group = G2,ymin = ymin,ymax = ymax),alpha = 0.2) + geom_line(aes(color = G2,y=y_pred))

mod2 = lm(y~G1*G2,data)
p = ggplot(data,aes(x=G1,y=y)) + geom_boxplot(aes(color = G2))
p + geom_smooth(aes(group = G2,color = G2),span = .3)
# p + stat_plotlm(fitted_model = mod2,aes(group = G2,color = G2,G1=G1,G2=G2),se=T)

predicted_values = data.frame(summary(lsmeans(mod2,~G1+G2,weights = 'cell')))
predicted_values$y = predicted_values$lsmean
p + geom_ribbon(data = predicted_values,aes(ymin = lower.CL,ymax = upper.CL,group = G2),alpha = 0.2) + geom_line(data = predicted_values,aes(group = G2,color = G2,y = lsmean))

p = ggplot(data,aes(x=xx,y=y)) + geom_point(aes(color = G2)) + facet_wrap(~G1,scales = 'free')
p + geom_smooth(aes(color = G2),method = lm,fullrange = F)

predicted_values = data.frame(summary(lsmeans(ref.grid(my_lm,with(data,list(G1=levels(G1),G2=levels(G2),xx = seq(min(xx),max(xx),length=100)))),~G1+G2+xx,weights = 'cell')))
predicted_values$y = predicted_values$lsmean
p + geom_ribbon(data = predicted_values,aes(ymin = lower.CL,ymax = upper.CL,group = G2),alpha = 0.2) + geom_line(data = predicted_values,aes(group = G2,color = G2,y = lsmean))



p + stat_plotlm(fitted_model = my_lm,aes(color = G2,G1=G1,G2=G2,xx=xx),fullrange = F,se=T)
p %+% data2 + geom_ribbon(aes(group = G2,ymin = ymin,ymax = ymax),alpha = 0.2) + geom_line(aes(color = G2,y=y_pred))
data = add_intervals(data,my_lm)
p = ggplot(data,aes(x=xx,y=y)) + geom_point(aes(color = G2)) + facet_wrap(~G1,scales = 'free')+ geom_ribbon(aes(group = G2,ymin = ymin,ymax = ymax),alpha = 0.2) + geom_line(aes(color = G2,y=y_pred))
p

mod3 = lm(y~G1*G2 + poly(xx,2),data)
p = ggplot(data,aes(x=xx,y=y)) + geom_point(aes(color = G2)) + facet_wrap(~G1,scales = 'free')
p + stat_plotlm(fitted_model = mod3,aes(color = G2,G1=G1,G2=G2,xx=xx),fullrange = F,se=T)
data = add_intervals(data,mod3)
p = ggplot(data,aes(x=xx,y=y)) + geom_point(aes(color = G2)) + facet_wrap(~G1,scales = 'free')+ geom_ribbon(aes(group = G2,ymin = ymin,ymax = ymax),alpha = 0.2) + geom_line(aes(color = G2,y=y_pred))


mod4 = lm(y~G1*G2,data)
p = ggplot(data,aes(x=G1,y=y)) + geom_point(aes(color = G2))
p + stat_plotlm(fitted_model = mod4,aes(group = G2,color = G2,G1=G1,G2=G2,xx=xx),fullrange = F,se=T)

library(lme4)
library(lmerTest)
mod5_lmer = lmer(y~G1*poly(xx,2) + (1|G2),data)
mod5_lm = lm(y~G1*poly(xx,2)+G2,data)
p = ggplot(data,aes(x=xx,y=y)) + geom_point(aes(color = G2)) + facet_wrap(~G1,scales = 'free')
p + stat_plotlm(fitted_model = mod5_lmer,aes(group = G2,color = G2,G1=G1,G2=G2,xx=xx),fullrange = F,se=T)
p + stat_plotlm(fitted_model = mod5_lm,aes(group = G2,color = G2,G1=G1,G2=G2,xx=xx),fullrange = F,se=T)