NumPy ndarray.argmin() Method

NumPy ndarray.argmin()

The numpy.ndarray.argmin() method returns the index of the minimum value in a NumPy array. It can operate across the entire array or along a specified axis.

Syntax

ndarray.argmin(axis=None, out=None, *, keepdims=False)

Parameters

Parameter	Type	Description
axis	None, int, or tuple of ints, optional	Axis along which to find the index of the minimum value. If None, the index is for the flattened array.
out	ndarray, optional	Alternative output array for storing the result. Must have the same shape as the expected output.
keepdims	bool, optional	If True, the reduced dimensions are kept as size one, allowing proper broadcasting.

Return Value

Returns an integer index if axis=None, or an array of indices if an axis is specified. The result represents the position of the first occurrence of the minimum value.

---

Examples

1. Finding the Index of the Minimum Element

In this example, we create a 1D array and find the index of the minimum value.

import numpy as np

arr = np.array([4, 2, 1, 7, 5])

index = arr.argmin()
print(index)

Output:

2

The minimum value in the array is 1, which is at index 2.

2. Using the axis Parameter in ndarray.argmin()

Here, we find the index of the minimum value along a specific axis.

import numpy as np

arr = np.array([[3, 1, 4],
                [2, 8, 5]])

index_axis0 = arr.argmin(axis=0)
print(index_axis0)

index_axis1 = arr.argmin(axis=1)
print(index_axis1)

Output:

[1 0 0]
[1 0]

For axis=0 (columns), the indices correspond to the row positions of the minimum values in each column.

For axis=1 (rows), the indices correspond to the column positions of the minimum values in each row.

3. Keeping Dimensions with keepdims=True in ndarray.argmin()

In this example, we use keepdims=True to retain the reduced axis as a dimension of size one.

import numpy as np

arr = np.array([[3, 1, 4],
                [2, 8, 5]])

index = arr.argmin(axis=1, keepdims=True)
print(index)

Output:

[[1]
 [0]]

By setting keepdims=True, the result retains the original number of dimensions, making it useful for broadcasting.