forked from kscience/kmath
merge PR
This commit is contained in:
parent
a0e9180db6
commit
51f084d28b
@ -30,20 +30,17 @@ fun main() {
|
||||
|
||||
println("Real alpha:\n$alpha")
|
||||
|
||||
// also take sample of size 20 from normal distribution for x TODO rename
|
||||
// also take sample of size 20 from normal distribution for x
|
||||
val x = randNormal(
|
||||
intArrayOf(20, 5),
|
||||
randSeed
|
||||
)
|
||||
|
||||
// calculate y and add gaussian noise (N(0, 0.05))
|
||||
// TODO: please add an intercept: Y = beta * X + alpha + N(0,0.5)
|
||||
val y = x dot alpha
|
||||
y += y.randNormalLike(randSeed) * 0.05
|
||||
|
||||
// now restore the coefficient vector with OSL estimator with SVD
|
||||
// TODO: you need to change accordingly [X 1] [alpha beta] = Y
|
||||
// TODO: inverting [X 1] via SVD
|
||||
val (u, singValues, v) = x.svd()
|
||||
|
||||
// we have to make sure the singular values of the matrix are not close to zero
|
||||
|
Loading…
Reference in New Issue
Block a user