Scribd
Upload
11 views3 pages

NumPy Basics for Scientific Computing

Uploaded by

SIVAVEDATHRI
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
11 views3 pages

NumPy Basics for Scientific Computing

Uploaded by

SIVAVEDATHRI
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

NumPy — Notes

1. Introduction
• NumPy (Numerical Python) is a core Python library for scientific computing.
• Provides support for large, multi-dimensional arrays and matrices.
• Offers mathematical functions, linear algebra, random number generation, and more.

2. NumPy Arrays
• ndarray: Core object representing homogeneous multidimensional data.

import numpy as np
arr = [Link]([1, 2, 3, 4])
print(arr)

Array creation methods

[Link]((2,3)) # 2x3 matrix of zeros


[Link]((2,2)) # 2x2 matrix of ones
[Link](0,10,2) # [0,2,4,6,8]
[Link](0,1,5) # 5 values between 0 and 1
[Link](3) # 3x3 Identity matrix

3. Array Attributes

arr = [Link](12).reshape(3,4)
[Link] # (3, 4)
[Link] # 2
[Link] # 12
[Link] # int64
[Link] # bytes per element

1
4. Indexing & Slicing

arr = [Link]([10,20,30,40,50])
arr[0] # 10
arr[1:4] # [20 30 40]

mat = [Link]([[1,2,3],[4,5,6]])
mat[0,1] # 2
mat[:,1] # column → [2 5]

5. Array Operations
• Vectorized arithmetic (faster than loops)

a = [Link]([1,2,3])
b = [Link]([4,5,6])
print(a + b) # [5 7 9]
print(a * 2) # [2 4 6]

• Universal functions (ufuncs)

[Link](a) # [1.0, 1.41, 1.73]


[Link](b)
[Link](a)

6. Linear Algebra

A = [Link]([[1,2],[3,4]])
B = [Link]([[5,6],[7,8]])

[Link](A,B) # matrix multiplication


[Link](A) # inverse
[Link](A) # eigenvalues & vectors

2
7. Random Module

[Link](3) # uniform [0,1)


[Link](3) # normal distribution
[Link](1,10,5) # 5 ints between 1-9

8. Broadcasting
• Different shapes are combined automatically when compatible.

A = [Link]((3,3))
B = [Link]([1,2,3])
print(A + B) # row-wise broadcast

9. File I/O

[Link]("[Link]", arr, delimiter=",")


arr2 = [Link]("[Link]", delimiter=",")

10. Use Cases


• Data preprocessing in ML/AI
• Efficient numerical computations
• Used with Pandas, SciPy, scikit-learn, TensorFlow, PyTorch

Conclusion
NumPy is the foundation of Python’s data science ecosystem, enabling fast mathematical operations and
serving as the base for higher-level libraries.

Common questions

Powered by AI

NumPy provides advanced random number generation capabilities compared to standard Python libraries, offering flexibility and a wide range of distributions. Functions such as 'np.random.rand' for uniform distribution, 'np.random.randn' for normal distribution, and 'np.random.randint' for generating random integers, allow for rigorous control over randomness in scientific simulations and data sampling tasks .

NumPy handles linear algebra operations using functions optimized for matrix calculations. Functions include 'np.dot' for matrix multiplication, 'np.linalg.inv' for computing the inverse of a matrix, and 'np.linalg.eig' for finding eigenvalues and eigenvectors. These functions allow for efficient execution of complex mathematical tasks, making NumPy a powerful tool for scientific computation .

NumPy serves as the foundational package for numerical computations in Python's data science ecosystem. It provides the core functionality needed for data manipulation and scientific calculations. NumPy's efficient array handling is utilized by libraries like Pandas for data analysis, SciPy for scientific computing, and machine learning frameworks like scikit-learn, TensorFlow, and PyTorch. These integrations allow for seamless operations and efficient data processing workflows across various domains .

Broadcasting in NumPy refers to the automatic extension of arrays to perform arithmetic operations on arrays of different shapes. This happens when the smaller array is 'broadcasted' across the larger array so that they have compatible shapes for vectorized operations. For example, adding a one-dimensional array B = [1, 2, 3] to a two-dimensional array A = np.ones((3,3)) results in broadcasting B across each row of A, producing a result where each row of A has B added to it .

NumPy provides support for large, multi-dimensional arrays and matrices through the 'ndarray' object, allowing for efficient data storage and manipulation. Key functions involved include array creation methods such as 'np.zeros', 'np.ones', 'np.arange', and 'np.linspace'. Operations on these arrays are optimized using vectorized arithmetic, universal functions like 'np.sqrt' and 'np.exp', and linear algebra operations including 'np.dot' for matrix multiplication and 'np.linalg.inv' for calculating inverses .

NumPy's file I/O capabilities are crucial for data processing as they allow for efficient reading and writing of large data sets to and from disk. Methods such as 'np.savetxt' and 'np.loadtxt' offer streamlined operations to save and load data in text format. These capabilities enable integration with other data processing tools and facilitate data sharing and persistence across different environments .

NumPy enhances performance by using vectorized operations, which perform element-wise calculations across entire arrays without the need for explicit loops. This reduces the overhead of iterating over elements in Python and takes advantage of optimized C and Fortran routines underneath, leading to significant speed improvements. Universal functions (ufuncs) further enhance performance by providing fast element-wise operations on arrays .

Indexing and slicing in NumPy allows for more advanced and efficient operations compared to standard Python lists. NumPy arrays support multi-dimensional indexing and slicing, which enables extraction of sub-arrays and specific elements with more precision. For example, 'arr[1:4]' extracts a slice from a one-dimensional array, and 'mat[:,1]' extracts an entire column from a two-dimensional array. These capabilities facilitate complex data manipulations that are computationally expensive with Python lists .

NumPy offers several methods for creating arrays, contributing to its flexibility. Some methods include 'np.zeros' for creating an array of zeros, 'np.ones' for an array of ones, 'np.arange' for a range of values, 'np.linspace' for evenly spaced numbers over a specified interval, and 'np.eye' for an identity matrix. These methods allow users to initialize arrays of different shapes and sizes efficiently, catering to a wide range of computational needs .

Array attributes in NumPy such as 'shape', 'ndim', 'size', 'dtype', and 'itemsize' provide essential information on the structure and characteristics of arrays. These attributes assist in ensuring appropriate data manipulation by allowing users to understand the dimensionality ('shape'), data type ('dtype'), and memory usage ('itemsize') of arrays. This understanding enables efficient planning of operations such as reshaping, slicing, and broadcasting, which are critical in scientific computing .

You might also like