1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166

use traits::operations::{Transpose, ApproxEq};
use traits::structure::{ColSlice, Eye, Indexable, Diag, SquareMat, BaseFloat};
use traits::geometry::Norm;
use std::cmp::min;
use std::ops::{Mul, Add, Sub};

/// Get the householder matrix corresponding to a reflexion to the hyperplane
/// defined by `vec`. It can be a reflexion contained in a subspace.
///
/// # Arguments
/// * `dim` - the dimension of the space the resulting matrix operates in
/// * `start` - the starting dimension of the subspace of the reflexion
/// * `vec` - the vector defining the reflection.
pub fn householder_matrix<N, V, M>(dim: usize, start: usize, vec: V) -> M
    where N: BaseFloat,
          M: Eye + Indexable<(usize, usize), N>,
          V: Indexable<usize, N> {
    let mut qk : M = Eye::new_identity(dim);
    let subdim = vec.shape();

    let stop = subdim + start;

    assert!(dim >= stop);

    for j in (start .. stop) {
        for i in (start .. stop) {
            unsafe {
                let vv = vec.unsafe_at(i - start) * vec.unsafe_at(j - start);
                let qkij = qk.unsafe_at((i, j));
                qk.unsafe_set((i, j), qkij - vv - vv);
            }
        }
    }
    qk
}

/// QR decomposition using Householder reflections.
///
/// # Arguments
/// * `m` - matrix to decompose
pub fn qr<N, V, M>(m: &M) -> (M, M)
    where N: BaseFloat,
          V: Indexable<usize, N> + Norm<N>,
          M: Copy + Eye + ColSlice<V> + Transpose + Indexable<(usize, usize), N> +
             Mul<M, Output = M> {
    let (rows, cols) = m.shape();
    assert!(rows >= cols);
    let mut q : M = Eye::new_identity(rows);
    let mut r = *m;

    for ite in 0..min(rows - 1, cols) {
        let mut v = r.col_slice(ite, ite, rows);
        let alpha =
            if unsafe { v.unsafe_at(ite) } >= ::zero() {
                -Norm::norm(&v)
            }
            else {
                Norm::norm(&v)
            };
        unsafe {
            let x = v.unsafe_at(0);
            v.unsafe_set(0, x - alpha);
        }
        if !::is_zero(&v.normalize_mut()) {
            let qk: M = householder_matrix(rows, ite, v);
            r = qk * r;
            q = q * Transpose::transpose(&qk);
        }
    }

    (q, r)
}

/// Eigendecomposition of a square matrix using the qr algorithm.
pub fn eigen_qr<N, V, VS, M>(m: &M, eps: &N, niter: usize) -> (M, V)
    where N:  BaseFloat,
          V:  Mul<M, Output = V>,
          VS: Indexable<usize, N> + Norm<N>,
          M:  Indexable<(usize, usize), N> + SquareMat<N, V> + Add<M, Output = M> +
              Sub<M, Output = M> + ColSlice<VS> +
              ApproxEq<N> + Copy {
    let mut eigenvectors: M = ::one::<M>();
    let mut eigenvalues = *m;
    // let mut shifter: M = Eye::new_identity(rows);

    let mut iter = 0;
    for _ in 0..niter {
        let mut stop = true;

        for j in 0..::dim::<M>() {
            for i in 0..j {
                if unsafe { eigenvalues.unsafe_at((i, j)) }.abs() >= *eps {
                    stop = false;
                    break;
                }
            }

            for i in j + 1..::dim::<M>() {
                if unsafe { eigenvalues.unsafe_at((i, j)) }.abs() >= *eps {
                    stop = false;
                    break;
                }
            }
        }

        if stop {
            break;
        }
        iter = iter + 1;

        let (q, r) = qr(&eigenvalues);;

        eigenvalues  = r * q;
        eigenvectors = eigenvectors * q;
    }

    (eigenvectors, eigenvalues.diag())
}

/// Cholesky decomposition G of a square symmetric positive definite matrix A, such that A = G * G^T
///
/// # Arguments
/// * `m` - square symmetric positive definite matrix to decompose
pub fn cholesky<N, V, VS, M>(m: &M) -> Result<M, &'static str>
    where N:  BaseFloat,
          V:  Mul<M, Output = V>,
          VS: Indexable<usize, N> + Norm<N>,
          M:  Indexable<(usize, usize), N> + SquareMat<N, V> + Add<M, Output = M> +
              Sub<M, Output = M> + ColSlice<VS> +
              ApproxEq<N> + Copy {

    let mut out = m.clone().transpose();

    if !ApproxEq::approx_eq(&out, &m) {
        return Err("Cholesky: Input matrix is not symmetric");
    }

    for i in 0 .. out.nrows() {
        for j in 0 .. (i + 1) {

            let mut sum: N = out[(i, j)];

            for k in 0 .. j {
                sum = sum - out[(i, k)] * out[(j, k)];
            }

            if i > j {
                out[(i, j)] = sum / out[(j, j)];
            }
            else if sum > N::zero() {
                out[(i, i)] = sum.sqrt();
            }
            else {
                return Err("Cholesky: Input matrix is not positive definite to machine precision");
            }
        }
    }

    for i in 0 .. out.nrows() {
        for j in i + 1 .. out.ncols() {
            out[(i, j)] = N::zero();
        }
    }

    return Ok(out);
}