MCPcopy Index your code
hub / github.com/numpy/numpy / percentile

Function percentile

numpy/lib/_function_base_impl.py:4065–4259  ·  view source on GitHub ↗

Compute the q-th percentile of the data along the specified axis. Returns the q-th percentile(s) of the array elements. Parameters ---------- a : array_like of real numbers Input array or object that can be converted to an array. q : array_like of float Per

(a,
               q,
               axis=None,
               out=None,
               overwrite_input=False,
               method="linear",
               keepdims=False,
               *,
               weights=None)

Source from the content-addressed store, hash-verified

4063
4064@array_function_dispatch(_percentile_dispatcher)
4065def percentile(a,
4066 q,
4067 axis=None,
4068 out=None,
4069 overwrite_input=False,
4070 method="linear",
4071 keepdims=False,
4072 *,
4073 weights=None):
4074 """
4075 Compute the q-th percentile of the data along the specified axis.
4076
4077 Returns the q-th percentile(s) of the array elements.
4078
4079 Parameters
4080 ----------
4081 a : array_like of real numbers
4082 Input array or object that can be converted to an array.
4083 q : array_like of float
4084 Percentage or sequence of percentages for the percentiles to compute.
4085 Values must be between 0 and 100 inclusive.
4086 axis : {int, tuple of int, None}, optional
4087 Axis or axes along which the percentiles are computed. The
4088 default is to compute the percentile(s) along a flattened
4089 version of the array.
4090 out : ndarray, optional
4091 Alternative output array in which to place the result. It must
4092 have the same shape and buffer length as the expected output,
4093 but the type (of the output) will be cast if necessary.
4094 overwrite_input : bool, optional
4095 If True, then allow the input array `a` to be modified by intermediate
4096 calculations, to save memory. In this case, the contents of the input
4097 `a` after this function completes is undefined.
4098 method : str, optional
4099 This parameter specifies the method to use for estimating the
4100 percentile. There are many different methods, some unique to NumPy.
4101 See the notes for explanation. The options sorted by their R type
4102 as summarized in the H&F paper [1]_ are:
4103
4104 1. 'inverted_cdf'
4105 2. 'averaged_inverted_cdf'
4106 3. 'closest_observation'
4107 4. 'interpolated_inverted_cdf'
4108 5. 'hazen'
4109 6. 'weibull'
4110 7. 'linear' (default)
4111 8. 'median_unbiased'
4112 9. 'normal_unbiased'
4113
4114 The first three methods are discontinuous. NumPy further defines the
4115 following discontinuous variations of the default 'linear' (7.) option:
4116
4117 * 'lower'
4118 * 'higher',
4119 * 'midpoint'
4120 * 'nearest'
4121
4122 .. versionchanged:: 1.22.0

Callers 2

test_percentile_outMethod · 0.85
test_result_valuesMethod · 0.85

Calls 4

_quantile_is_validFunction · 0.85
_weights_are_validFunction · 0.85
_quantile_uncheckedFunction · 0.85
anyMethod · 0.45

Tested by 2

test_percentile_outMethod · 0.68
test_result_valuesMethod · 0.68

Used in the wild real call sites across dependent graphs

searching dependent graphs…