| | |
- Probability
-
- Affine
class Affine(Probability) |
| |
the probability law governing Z*s + m, if Z obeys probBase.
|
| |
Methods defined here:
- Affine(self, xZ)
- Density(self, x)
- Distribution(self, x)
- InverseDistribution(self, y)
- Sample(self)
- Standardize(self, xX)
- __init__(self, probBase, xM=0.0, xS=1.0)
- __repr__(self)
- getMean(self)
- getVariance(self)
Methods inherited from Probability:
- DensityTest(self, xX, **dArgs)
- Compare the Density() with the numerically differentiated Distribution().
This doesn't make much sense in the abstract base class. But in general, the Density() method will be redefined, and this will be a real test.
- InverseTest(self, xY, **dArgs)
- Run a y value through InverseDistribution and then through Distribution, and compare the results.
- __str__(self)
- getIQR(self)
- getMedian(self)
- getStDev(self)
- html(self)
|
class Probability |
| |
the base class for probability laws
|
| |
Methods defined here:
- Density(self, xX, xDelta=4.9999999999999998e-07)
- the density function of the probability law, equal to the derivative of the distribution function
By default, numerically differentiate the distribution function.
- DensityTest(self, xX, **dArgs)
- Compare the Density() with the numerically differentiated Distribution().
This doesn't make much sense in the abstract base class. But in general, the Density() method will be redefined, and this will be a real test.
- Distribution(self, xX)
- the distribution function of the probability law
- InverseDistribution(self, xY, xX=0.0, xEpsilon=9.9999999999999995e-07)
- the inverse of the distribution function of the probability law
By default, invert using Newton's method.
Note that using a numerically calculated derivative, as in the default Density(), with Newton's method as here, is numerically problematic. One or the other, or both, should be redefined in any (non-abstract) extension. Fortunately, alternate expressions are available for commonly-used distributions.
- InverseTest(self, xY, **dArgs)
- Run a y value through InverseDistribution and then through Distribution, and compare the results.
- Sample(self)
- Return a random sample distributed according to the probability law.
By default, pass a uniform random sample to InverseDistribution().
Similarly, passing a sample through Distribution() will give you a uniform sample on [0, 1].
- __repr__(self)
- __str__(self)
- getIQR(self)
- getMean(self)
- getMedian(self)
- getStDev(self)
- getVariance(self)
- html(self)
| |