tf.keras.ops.nanpercentile

Compute the q-th percentile(s) of the data along the specified axis, while ignoring NaNs.

tf.keras.ops.nanpercentile(
    x, q, axis=None, method='linear', keepdims=False
)

x Input tensor.
q Percentile or sequence of percentiles to compute. Values must be between 0 and 100 inclusive.
axis Axis or axes along which the percentiles are computed.
method A string specifies the method to use for estimating the percentile. Available methods are "linear", "lower", "higher", "midpoint", and "nearest". Defaults to "linear". If the desired percentile lies between two data points i < j:

  • "linear": i + (j - i) * fraction, where fraction is the fractional part of the index surrounded by i and j.
  • "lower": i.
  • "higher": j.
  • "midpoint": (i + j) / 2
  • "nearest": i or j, whichever is nearest.
keepdims If True, reduced axes are kept with size 1.

The percentile(s) ignoring NaNs.

Examples:

import keras
from keras import ops
x = keras.ops.array([1., 2., 3., 4.])
keras.ops.nanpercentile(x, 50)
2.5
x = keras.ops.array([1., 2., float("nan"), 4.])
keras.ops.nanpercentile(x, 50)
2.0
x = keras.ops.array([1., 2., 3., 4.])
keras.ops.nanpercentile(x, [25, 75])
array([1.75, 3.25])
x = keras.ops.array([[1., 2., float("nan")],
                     [4., 5., 6.]])
keras.ops.nanpercentile(x, 50, axis=1)
array([1.5, 5.0])