Parallel Computing for Data Science

3.4. EXAMPLE: ALL POSSIBLE REGRESSIONS 45

3.4.3 Sample Run

Here is some sample output:

> snowtest(c8,100,4,2)

[,1] [,2] [,3] [,4] [,5]

[1,] 0.21941625 1 0 0 0

[2,] 0.05960716 2 0 0 0

[3,] 0.11090411 3 0 0 0

[4,] 0.15092073 4 0 0 0

[5,] 0.26576805 1 2 0 0

[6,] 0.35730378 1 3 0 0

[7,] 0.32840075 1 4 0 0

[8,] 0.17534962 2 3 0 0

[9,] 0.20841665 2 4 0 0

[10,] 0.27900555 3 4 0 0

Here simulated data of size n = 100 was generated, with p = 4 predictors

and a maximum predictor set size of k = 2. The highest adjusted R

2

value

was about 0.36, for the model using predictors 1 and 3, i.e., columns 1 and

3 of x.

3.4.4 Code Analysis

As noted in Section 3.3, parallel code does ...