# FUNCTION

# Public

# RUN (&OPTIONAL INTERACTIVE?)

Run all tests in the test suite.

# Private

# APPROXIMATION-ERROR (F F-APPROX INTERVAL &OPTIONAL (N-GRID 1000))

Approximation error, using MAXIMUM-ON-GRID.

# ASSERT-DISTRIBUTIVE-CONVERT-OP (A B CONVERT OP)

Check that OP distributes over CONVERT.

# ASSERT-NOT-RELATION (RELATION &REST ARGUMENT-LISTS)

Assert that RELATION does not hold, called with each set of arguments.

# ASSERT-RELATION (RELATION &REST ARGUMENT-LISTS)

Assert RELATION called with each set of arguments.

# DO-MATRIX-CONVERT-OPS (TEST CONVERTS &KEY (OPS (LIST #'E+ #'E- #'E*)))

Funcall TEST with CONVERT and each operation in OPs.

# MAXIMUM-ON-GRID (F INTERVAL &OPTIONAL (N-GRID 1000))

Maximum of F on a grid of N-GRID equidistand points in INTERVAL.

# NUM=-TEST-STRUCT-A (INSTANCE)

@arg[extid]{A @class{extid}}
@return[sytemid]{puri:uri or nil}
Returns the System ID part of this External ID.

# NUM=-TEST-STRUCT-B (INSTANCE)

@arg[extid]{A @class{extid}}
@return[sytemid]{puri:uri or nil}
Returns the System ID part of this External ID.

# PRECISE-CENTRAL-MOMENTS (SEQUENCE)

First 4 central moments, calculated using rationals, returned as four values, normalized by the length of the sequence.
Slow, but useful for testing as it does not suffer from approximation error.

# PRECISE-WEIGHTED-CENTRAL-MOMENTS (SEQUENCE WEIGHTS)

First 4 weighted central moments, calculated using rationals, returned as four values, normalized by the length of the sequence.
Slow, but useful for testing as it does not suffer from approximation error.

# RANDOM-FLOATS (N MEAN &OPTIONAL (ELEMENT-TYPE 'DOUBLE-FLOAT))

Return a N-element vector of random floats (with given ELEMENT-TYPE). A uniform random number from either [-1,0] or [0,3] (with equal probability) is added to MEAN, which ensures nonzero third and fourth central moments. Higher abolute value of MEAN makes the calculation of higher central moments more ill-conditioned when using floats.

# RANDOM-WEIGHTS (N RANGE &OPTIONAL (ELEMENT-TYPE 'DOUBLE-FLOAT))

Random weights between (exp 1) and (exp (1+ range)), with given element-type.

# TEST-MOMENTS (N MEAN &OPTIONAL (ELEMENT-TYPE 'DOUBLE-FLOAT))

Test that moments calculated precisely and with accumulators are equal.

# TEST-WEIGHTED-MOMENTS (N MEAN &KEY (WEIGHT-RANGE 4) (ELEMENT-TYPE 'DOUBLE-FLOAT))

Test that moments calculated precisely and with accumulators are equal.

# Undocumented

# ASSERT-NOT-PAIRED-RELATION (RELATION1 RELATION2 &REST ARGUMENT-LISTS)

# ASSERT-PAIRED-RELATION (RELATION1 RELATION2 &REST ARGUMENT-LISTS)

# ASSERT-RELATION-CORNER-CASES (&REST RELATIONS)

# COPY-NUM=-TEST-STRUCT (INSTANCE)

# MAKE-NUM=-TEST-STRUCT (&KEY ((A DUM339) NIL) ((B DUM340) NIL))

# SETFNUM=-TEST-STRUCT-A (NEW-VALUE INSTANCE)

# SETFNUM=-TEST-STRUCT-B (NEW-VALUE INSTANCE)

# NUM=-TEST-STRUCT-P (OBJECT)

# TEST-CHEBYSHEV-APPROXIMATE (F INTERVAL N-POLYNOMIALS TEST-INTERVAL &REST REST)

# TEST-POOLED-MOMENTS (N MEAN &OPTIONAL (ELEMENT-TYPE 'DOUBLE-FLOAT))

# GENERIC-FUNCTION

# Private

# MEAN (OBJECT &KEY WEIGHTS)

The mean of elements in OBJECT.

# MEDIAN (OBJECT)

Median of OBJECT.

# VARIANCE (OBJECT &KEY WEIGHTS)

Variance of OBJECT. For samples, normalized by the weight-1 (and thus unbiased if certain assumptions hold, eg weights that count frequencies).

# CLASS

# Private

# NUM=-TEST-STRUCT

Structure for testing DEFINE-STRUCTURE-num=.