Numpy Argmax in Two Dimensions

Numpy is a fundamental package for scientific computing in Python. It provides a high-performance multidimensional array object, and tools for working with these arrays. One of the useful functions provided by Numpy is argmax, which returns the indices of the maximum values along an axis. When working with two-dimensional arrays, understanding how to use argmax effectively can be incredibly useful for data analysis and manipulation.

In this article, we will explore how to use the argmax function in two dimensions, providing detailed examples and explanations to help you master this functionality.

Understanding `numpy.argmax`

The numpy.argmax function returns the indices of the maximum values along a specified axis. In a two-dimensional array, you can specify whether you want to find the maximum values in rows or columns by setting the axis parameter.

Syntax of `numpy.argmax`

numpy.argmax(a, axis=None, out=None)

a: Input array.
axis: By default, the index is into the flattened array, otherwise along the specified axis.
out: If provided, the result will be inserted into this array. It should be of the appropriate shape and dtype.

Example 1: Basic Usage of `argmax` in a 2D Array

import numpy as np

array_2d = np.array([[1, 3, 5], [6, 2, 8], [7, 9, 4]])
result = np.argmax(array_2d, axis=1)
print(result)  # Output will be [2, 2, 1]

Output:

Example 2: Using `argmax` Without Specifying Axis

import numpy as np

array_2d = np.array([[1, 3, 5], [6, 2, 8], [7, 9, 4]])
result = np.argmax(array_2d)
print(result)  # Output will be 5

Output:

Detailed Examples with `numpy.argmax`

Let’s dive deeper with more examples to understand the usage of numpy.argmax in various scenarios with two-dimensional arrays.

Example 3: Finding Column Indices of Maximum Values

import numpy as np

array_2d = np.array([[10, 20, 30], [15, 25, 5], [35, 5, 15]])
result = np.argmax(array_2d, axis=0)
print(result)  # Output will be [2, 1, 0]

Output:

Example 4: Using `argmax` with a Flattened Array

import numpy as np

array_2d = np.array([[10, 20, 30], [15, 25, 5], [35, 5, 15]])
result = np.argmax(array_2d.flatten())
print(result)  # Output will be 6

Output:

Example 5: Specifying Output Array for Results

import numpy as np

array_2d = np.array([[10, 20, 30], [15, 25, 5], [35, 5, 15]])
output_array = np.empty((3,), dtype=int)
np.argmax(array_2d, axis=1, out=output_array)
print(output_array)  # Output will be [2, 1, 0]

Output:

Example 6: Argmax with Randomly Generated Array

import numpy as np

np.random.seed(0)
array_2d = np.random.randint(1, 100, (3, 3))
result = np.argmax(array_2d, axis=1)
print(result)  # Output will be [2, 2, 0]

Output:

Example 7: Argmax in a Real-World Dataset

import numpy as np

# Simulating a dataset where rows are samples and columns are features
data = np.random.rand(10, 5) * 100
# Finding the feature (column) with the maximum value for each sample (row)
max_feature_indices = np.argmax(data, axis=1)
print(max_feature_indices)  # Output will be indices of the max feature per sample

Output:

Example 8: Handling Higher Dimensional Data

import numpy as np

data = np.random.rand(2, 3, 4) * 100
# Applying argmax on the last axis of a 3D array
max_indices = np.argmax(data, axis=-1)
print(max_indices)  # Output will be a 2x3 array

Output:

Example 9: Using Argmax with Structured Arrays

import numpy as np

dtype = [('name', 'S10'), ('height', float), ('age', int)]
values = [('Alice', 165.3, 34), ('Bob', 175.0, 45), ('Charlie', 155.0, 25)]
structured_array = np.array(values, dtype=dtype)
max_index = np.argmax(structured_array['height'])
print(max_index)  # Output will be 1

Output:

Example 10: Argmax with NaN Values

import numpy as np

array_2d = np.array([[np.nan, 2, np.nan], [3, np.nan, 1], [4, 2, np.nan]])
result = np.nanargmax(array_2d, axis=1)
print(result)  # Output will be [1, 0, 0]

Output: