/*
* Assumes the data is gaussian, rejects outliers that are nsigma away from the mean.
* Sets outlier values to be NAN in the input data array.
* returns the number of rejected values.
*/
int reject_outliers(double data[], int size, float nsigma)
{
double mean, sigma;
int i, reject_count;
mean = compute_mean(data, 0,size);
sigma = compute_sigma(data, size, mean);
reject_count = 0;
for (i=0; i<size; i++) {
if (isnan(data[i])) continue;
if (fabs(mean - data[i]) > nsigma*sigma) {
data[i] = NAN;
reject_count++;
}
}
return reject_count;
}
void householders(
orthotope<T> const& A
)
{
BOOST_ASSERT(2 == A.order());
BOOST_ASSERT(A.hypercube());
std::size_t const n = A.extent(0);
orthotope<T> R = A.copy();
orthotope<T> Q({n, n});
for (std::size_t l = 0; l < n; ++l)
Q(l, l) = 1.0;
for (std::size_t l = 0; l < (n - 1); ++l)
{
T const sigma = compute_sigma(R, n, l);
boost::int16_t const sign = compute_sign(R(l, l));
#if defined(HPXLA_DEBUG_HOUSEHOLDERS)
std::cout << (std::string(80, '#') + "\n")
<< "ROUND " << l << "\n\n";
print(sigma, "sigma");
print(sign, "sign");
#endif
orthotope<T> w({n});
w(l) = R(l, l) + sign * sigma;
for (std::size_t i = (l + 1); i < w.extent(0); ++i)
w(i) = R(i, l);
#if defined(HPXLA_DEBUG_HOUSEHOLDERS)
print(w, "u");
#endif
T const w_norm = euclidean_norm(w);
for (std::size_t i = l; i < n; ++i)
w(i) /= w_norm;
#if defined(HPXLA_DEBUG_HOUSEHOLDERS)
print(w, "v");
#endif
orthotope<T> H = compute_H(w);
#if defined(HPXLA_DEBUG_HOUSEHOLDERS)
print(H, "H");
#endif
R = matrix_multiply(H, R);
Q = matrix_multiply(Q, H);
for (std::size_t i = l + 1; i < n; ++i)
R(i, l) = 0;
}
#if defined(HPXLA_DEBUG_HOUSEHOLDERS)
std::cout << std::string(80, '#') << "\n";
#endif
print(A, "A");
print(Q, "Q");
print(R, "R");
#if defined(HPXLA_DEBUG_HOUSEHOLDERS)
check_QR(A, Q, R);
#endif
}
