43 lines
1001 B
Julia
43 lines
1001 B
Julia
|
using CSV
|
||
|
|
using DataFrames
|
||
|
|
using LinearAlgebra
|
||
|
|
|
||
|
|
@doc raw"""
|
||
|
|
get_dataset(file_path::String, lambda::Float64)
|
||
|
|
|
||
|
|
Reads the ML-CUP-23/24 dataset from the given file path and returns the
|
||
|
|
necessary matrices and vectors to solve the problem. The LS problem is
|
||
|
|
solved with the standard solver in Julia.
|
||
|
|
|
||
|
|
### Input
|
||
|
|
|
||
|
|
- 'file_path' -- the path to the dataset file.
|
||
|
|
- 'lambda' -- the regularization parameter.
|
||
|
|
|
||
|
|
### Output
|
||
|
|
|
||
|
|
A tuple containing
|
||
|
|
|
||
|
|
- 'X_hat' -- the augmented matrix X.
|
||
|
|
- 'y_hat' -- the augmented vector y.
|
||
|
|
- 'start' -- the initial starting point.
|
||
|
|
- 'w_star' -- the optimal solution.
|
||
|
|
|
||
|
|
"""
|
||
|
|
function get_dataset(filePath::String, λ::Real)
|
||
|
|
X = CSV.File(filePath; header=false) |> DataFrame
|
||
|
|
X = Matrix(X)
|
||
|
|
|
||
|
|
m, n = size(X)
|
||
|
|
X_hat = [transpose(X); λ .* I(m)]
|
||
|
|
y_hat = [(rand(n).*2 .-1); zeros(m)]
|
||
|
|
|
||
|
|
# initial starting point
|
||
|
|
start = ones(m)
|
||
|
|
|
||
|
|
# w_star is the optimal solution provided by the julia solver
|
||
|
|
w_star = X_hat \ y_hat
|
||
|
|
|
||
|
|
return X_hat, y_hat, start, w_star
|
||
|
|
end
|