Seen returns the number of documents ever classified in the lifetime of this classifier.

Select one from pop

SqrtBig returns floor(sqrt(n)). It panics on n < 0.

* Creates a new matrix [A;B], where A is above B. */

* Finds the sum of two matrices. */

Interchanges two vectors (X <-> Y). ARGUMENTS X float or complex matrix Y float or complex matrix. Must have the same type as X. OP

ToBase produces n in base b. For example ToBase(2047, 22) -> [1, 5, 4] 1 * 22^0 1 5 * 22^1 110 4 * 22^2 1936

Update Y with 2 columns of A

Update Y with 2 columns of A

Update Y one cell with A[i:]*X

* * Applies a real elementary reflector H to a real m by n matrix A, * from either the left or the right. H is represented in the form * * H

* * Apply diagonal pivot (row and column swapped) to symmetric matrix blocks. * AR[0,0] is on diagonal and AL is block to the left of diagonal and A

Checks if this sweep event is below point p.

* Returns the Nesterov-Todd scaling W at points s and z, and stores the scaled variable in lmbda. W * z = W^{-T} * s = lmbda. W is a

consonant returns whether the letter at offset is a consonant

Here problem is already translated to epigraph format except original convex problem. We wrap it and create special CP epigraph kktsolver.

Calculates X.T*Y

For symmetric band matrix and vector X solve Y = alpha * A * X + beta * Y

For symmetric matrix and vector X solve Y = alpha * A * X + beta * Y

* The inverse product x := (y o\ x), when the 's' components of y are diagonal. */

The product x := y o y. The 's' components of y are diagonal and only the diagonals of x and y are stored.

Build Q in place by applying elementary reflectors in reverse order to an implied identity matrix. This forms Q = H(1)H(2) ... H(k) this is compatib

* * Unblocked QR decomposition with block reflector T. */

* The vector x is an element of S, with the 's' components stored in unpacked storage and off-diagonal entries scaled by sqrt(2). On return,

* Updates the Nesterov-Todd scaling matrix W and the scaled variable lmbda so that on exit W * zt = W^{-T} * st = lmbda. On entry,

vowel returns whether the letter at offset is a vowel

Parses a string representation of a BMN-Coordinate into a struct holding a BMN coordinate value. The reference ellipsoid of BMN coordinates is always

Make a copy C of A and compute C += alpha for all elements in the matrix if list of indexes is empty. Otherwise compute C[i] += alpha for indexes in c

Add is a convenience method for appending a contour to a polygon.

Compute in-place A[indexes[i]] += values[i]. Indexes are in column-major order.

* Add another matrix to this one in place, optimized for sparsity. */

Compute A = fn(A, x) by applying function fn element wise to A. For all i in indexes: A[indexes[i]] = fn(A[indexes[i]], values[i])

* Returns the contents of this matrix stored into a flat array (row-major). */

Get best values. Before start steps it takes nil.

BigratToFloat converts a *big.Rat to a float64 (with loss of precision).

BitLen returns the bit width of the non zero part of n.

CheckCode generates check code of chinese identity card number.

* Equivalent to AxP, but streamlined to take advantage of the datastructures. */

Return the flat column-major element array.

Solves a pair of primal and dual convex quadratic cone programs minimize (1/2)*x'*P*x + q'*x subject to G*x + s = h A*x = b s >= 0 maximize -

* Correlation() Calculate Pearson correlation = cov(X, Y) / (sigma_X * sigma_Y) This routine efficiently computes the correlation in one pass of the

Solves a convex optimization problem with a linear objective minimize f0(x) subject to fk(x) <= 0, k = 1, ..., mnl G*x <= h A*x = b. f is v

Cross product of two vectors. Returns error when ndim of either vector != 3.

Cross product with another vector, in-place. Returns error when ndim of either vector != 3.

return: alpha * X * Y

A = A + alpha * x * y.T; A is M*N, x is M*1, Y is N*1, 0 < R < E <= M, 0 < S < L <= N blas SYR2; blocked version

A = A + alpha * x * y.T; A is M*N, x is M*1, Y is N*1, 0 < R < E <= M, 0 < S < L <= N blas SYR; blocked version

generic triangular matrix rank update; A = A + alpha*X*Y.T

Absolute sum: \sum |X_i|

Copy the elements of the vectors X and Y.

* * Compute an LDLT factorization of a symmetric N-by-N matrix without pivoting. * * Arguments: * A On entry, the N-by-N matrix to be facto

* Convert this PivotMatrix into a DenseMatrix. */

returns a copy of the diagonal (not a slice)

## Transverse Mercator Projection Direct transverse mercator projection: Projection of an ellipsoid onto the surface of of a cylinder. Also known as G

Return the square of the Mahalanobis distance points.Rows() = dimensions. points.Cols() = number of points. target.Cols = 1

Euclidean norm: ||X||_2 = \sqrt {\sum X_i^2}

Apply a Givens rotation (X', Y') = (c X + s Y, c Y - s X) to the vectors X, Y

Exchange the elements of the vectors X and Y.

Equals returns true if both p1 and p2 describe exactly the same point.

Get the evaluated value by target function.

Create a new matrix from a list of matrices. New matrix has dimension (M, colmax) if direction is StackDown, and (rowmax, N) if direction is StackRigh