example/matrixpls.prediction-example2.R

In mronkko/matrixpls: Matrix-Based Partial Least Squares Estimation

library(lavaan)

## The industrialization and Political Democracy example 
## Bollen (1989), page 332. (Adopted from the lavaan package.)

model <- ' 
  # latent variable definitions
     ind60 =~ x1 + x2 + x3
     dem60 =~ y1 + a*y2 + b*y3 + c*y4
     dem65 =~ y5 + a*y6 + b*y7 + c*y8

  # regressions
    dem60 ~ ind60
    dem65 ~ ind60 + dem60

  # residual correlations
    y1 ~~ y5
    y2 ~~ y4 + y6
    y3 ~~ y7
    y4 ~~ y8
    y6 ~~ y8
'

matrixpls.out <- matrixpls(cov(PoliticalDemocracy),model)

# Calculate within-sample predictions

predictions <- predict(matrixpls.out, PoliticalDemocracy)

# Calculate root mean squared prediction errors

sqrt(apply((predictions-PoliticalDemocracy[61:75,])**2,2,mean))

# Q2 predictive relevance

q2(PoliticalDemocracy, predictions, model)


# Crossvalidation predictions using holdout samples and blindfolding

predictions.holdout <- matrixpls.crossvalidate(PoliticalDemocracy,
                                               model = model)

q2(PoliticalDemocracy, predictions.holdout, model)

predictions.blindfold <- matrixpls.crossvalidate(PoliticalDemocracy,
                                                 model = model,
                                                 blindfold = TRUE)

q2(PoliticalDemocracy, predictions.blindfold, model)