[frames] | no frames]

# Module weighted_percentile

source code

This computes order statistics on data with weights.

Functions
[ `float`, ... ]
 wp(data, wt, percentiles) Compute weighted percentiles. source code
`float`
 wtd_median(data, wt) The weighted median is the point where half the weight is above and half the weight is below. source code
`numpy.ndarray`
 wtd_median_across(list_of_vectors, wt) Takes a weighted component-by-component median of a sequence of vectors. source code

 test_wp() source code

 test_median() source code

 test() source code
 Variables __package__ = `'gmisclib'`

Imports: numpy

 Function Details

### wp(data, wt, percentiles)

source code

Compute weighted percentiles. If the weights are equal, this is the same as normal percentiles. Elements of the `data` and `wt` arrays correspond to each other and must have equal length (unless `wt` is `None`).

Parameters:
• `data` (A `numpy.ndarray` array or a `list` of numbers.) - The data.
• `wt` (`None` or a `numpy.ndarray` array or a `list` of numbers. All the weights must be non-negative and the sum must be greater than zero.) - How important is a given piece of data.
• `percentiles` (a `list` of numbers between 0 and 1.) - what percentiles to use. (Not really percentiles, as the range is 0-1 rather than 0-100.)
Returns: [ `float`, ... ]
the weighted percentiles of the data.

### wtd_median(data, wt)

source code

The weighted median is the point where half the weight is above and half the weight is below. If the weights are equal, this is the same as the median. Elements of the `data` and `wt` arrays correspond to each other and must have equal length (unless `wt` is `None`).

Parameters:
• `data` (A `numpy.ndarray` array or a `list` of numbers.) - The data.
• `wt` (`None` or a `numpy.ndarray` array or a `list` of numbers. All the weights must be non-negative and the sum must be greater than zero.) - How important is a given piece of data.
Returns: `float`
the weighted median of the data.

### wtd_median_across(list_of_vectors, wt)

source code

Takes a weighted component-by-component median of a sequence of vectors.

Parameters:
• `list_of_vectors` (any sequence of lists or numpy.ndarray. All the inside lists must be of the same length.) - the data to be combined
• `wt` (a vector of weights (one weight for each input vector) or None.) - sequence of numbers or None
Returns: `numpy.ndarray`
the median vector.

 Generated by Epydoc 3.0.1 on Thu Sep 22 04:25:02 2011 http://epydoc.sourceforge.net