R Code for Canonical Correlation Analysis

download R Code for Canonical Correlation Analysis

of 10

Transcript of R Code for Canonical Correlation Analysis

  • 8/14/2019 R Code for Canonical Correlation Analysis

    1/10

    ##############################################

    # R Code for canonical correlation analysis #

    ##############################################

    # We will use the built-in iris data set.

    # We will consider the entire data set (all three species)

    attach(iris)

    # We will standardize the variables first

    # by dividing by each column's standard deviation:

    # (we will remove column 5, the species labels)

    iris.std

  • 8/14/2019 R Code for Canonical Correlation Analysis

    2/10

    # u1 = 0.92*Sepal.Length - 0.39*Sepal.Width

    # v1 = 0.94*Petal.Length - 0.33*Petal.Width

    # Plotting the first set of canonical variables:

    u1

  • 8/14/2019 R Code for Canonical Correlation Analysis

    3/10

    colnames(xrels)

  • 8/14/2019 R Code for Canonical Correlation Analysis

    4/10

    .098,.308,-.27,-.183),

    ncol=4, byrow=T)

    R21

  • 8/14/2019 R Code for Canonical Correlation Analysis

    5/10

    P.value

  • 8/14/2019 R Code for Canonical Correlation Analysis

    6/10

    #### Testing about x2 in the model:

    ## Full model:

    my.n

  • 8/14/2019 R Code for Canonical Correlation Analysis

    7/10

    pt.est

    # A point prediction for this site is:

    # predicted CPU time = 151.8 hours, predicted disk input/output capacity

    = 349.6.

    middle.mat

  • 8/14/2019 R Code for Canonical Correlation Analysis

    8/10

    my.ys

  • 8/14/2019 R Code for Canonical Correlation Analysis

    9/10

    # Recall the variables:

    # X1 = Sales growth

    # X2 = Sales profitability

    # X3 = New account sales

    # X4 = Creativity Test

    # X5 = Mechanical Reasoning Test

    # X6 = Abstract Reasoning Test# X7 = Mathematics Test

    # X8 = Historical Facts Test

    # X9 = Sports Trivia Test

    # X10 = Music Trivia Test

    # We will try to predict the "performance" variables (x1, x2, x3) using

    the

    # "test score" variables (x4, x5, ..., x10).

    # Fitting the multivariate linear regression model:

    sales.mod.1

  • 8/14/2019 R Code for Canonical Correlation Analysis

    10/10

    X.mat.redu