NumPy

                    Numpy in Python

NumPy means Numeric Python. It is a package in Python which deals with numeric computation and processing of arrays. NumPy makes easy for us to do the numeric computations. Numpy plays a important role in Machine Learning. There are many inbuilt functions in Numpy, you just need to know when and how to use them. Numpy deals with different ranges of int and float data types. Now lets see how to use NumPy and do the mathematical computations.
Before using NumPy we have to import it. Generally we use alias while importing the library but it is only optional.

Creating Arrays:
import numpy as np
a1 = np.array([2, 1, 3, 4])   # single dimension array
print(a1)

Output: [2  1  3  4]

a2 = np.array([[1, 2, 1], [2, 1, 2], [1, 2, 3]])  # multi-dimension array
print(a2)

Output: [[1  2  1]
                  [2  1  2]
                  [1  2  3]]

By using the same procedure we can make the array of required dimensions.

arange() :
In arange() function we have to give the range between numbers and the interval and it will create the array of that range. Ending range value is exclusive. If we don't specify the interval between values then be default it will be 1. In function final range value is compulsory. By default initial range value is 0.

a3 = np.arange(10, 50, 10)
print(a3)

Output: [10  20  30  40]

linspace():
In linspace() function it needs three values, starting range value, ending range value which is inclusive here and the number of values you need within the range.

a4 = np.linspace(2, 6, 3)   # it means divide the range from 2 to 6 in 3 equal parts.
print(a4) 

Output: [2  4  6]

ones() and zeros():
ones() function will create the array consisting of value 1 only and similarly zeros() will create the array of zeros. These functions will need the dimensions of array we need.

a5 = np.ones((2, 2, 2))
print(a5)

Output: [ [ [ 1  1 ]
                      [ 1  1 ] ]
                   [ [ 1   1 ]
                      [ 1   1 ] ] ]

There are many more functions like these for example empty() will create the array of random values of given dimensions and eye() will create the identity matrix (I hope you know about identity matrix). 

Dimensions and Shape of matrix:
ndim gives the dimensions of array and shape will give the shape of array and size will give the size of array. Now what is the difference among three? Dimensions means 2D or 3D. Shape means the number of rows and columns. Size means total number of elements in the array.

print(a5.ndim)    # 3
print(a5.shape)   # (2, 2, 2)
print(a5.size)      # 8

If we want we can reshape our array. But we need to be careful with the dimensions. For example we have our 1D array above that we want to reshape into shape (2, 2).

print(a1.reshape(2, 2))

Output: [[2  1]
                  [3  4]]

We can use reshape function with arange and linspace to get the array in specified dimensions.

There are so many functions that we can use in Numpy like max() to get the maximum from array, min() to get the minimum from array, sum() to get the total sum of elements from array,  cumsum() whic prints the value equal to the sum of all previous elements in array. There are so many other functions also like sin(), tan(), log() etc. Use the numpy and explore these functions.

Dot product:
To perform the multiplication between two matrix we can use dot() function.

a = np.array([[1, 1], [2, 2], [3, 3]])
b = np.array([[1, 2, 3, 4, 5], [1, 2, 3, 4, 5]])
print(a.dot(b))

Output: [[2  4  6  8  10]
                  [4  8  12  16  20]
                  [6  12  18  24  30]]

We can use dot product in this way too-
np.dot(a, b)

Numpy is a cool library. Go and explore!


Comments

Post a Comment

Popular posts from this blog

Python Dictionary

                            Dictionary A dicitonary is a key-value pair collection which is unordered and can be changed or altered. We can not perform indexing on dictionary like lists but dictionary is still indexed. How? Well, we can use key inplace of index to get the value. We use curly braces { } to represent the dictionary.  E.g. dict1 = {'a':1, 'b':2, 'c':3, 'd':4} The value on left side of colon is called key, and value on right side of colon is called value. Hence key-value pair. We can use any datatype for keys and values. We can also use lists, sets, dictionary,etc. To access the value, we use keys. We can also use get() function for access the value. Just put the key in get() and value is accessed. E.g.   dict1['a']          # output : 1      dict1['c']          # output : 3     dict1.get('b')    ...
Read more

Loops in Python

                            Loops While programming sometimes there are conditions where we need to execute a block of code to execute several times and loops makes this task easy for us. Every programming language supports loops. Generally there are three types of loops i.e. for, while and do-while. But in Python we have only two types of loops while and for. In loops we have a condition and block of code. We execute the code till the condition satisfies. Loops can be used in netsed forms also. While loop :  In while loop we use the while keyword and code block runs till the condition satisfies. syntax - while condition:                  statements E.g. i = 0      while i < 10:         print(i, end = " ")         i += 1          Output : 0 1 2 3 4 5 6 7 8 9 In ab...
Read more

Python OOPs

  Object Oriented Programming in Python       Python is objected oriented programming. We can use classes and objects in Python. Class can be usderstood as a type of data structure which describes the properties and attributed of the object. Class is the blueprint of object. Object can be understood as the real world entity which has its own behaviour and state and instance of the class. Principles of OOP - Inheritance : In this concept we inherit the properties of one class to another. The class from which we inherit the property is called as parent class and other is child class. By inherit we means that we can use the property of one class to another. No need to write the code again. Polymorphism : Polymorphism means many forms. In this concept we means that we can have multiple functions of same name and can be differentiate on the basis of parameters. Encapsulation : In this concept, we encapsulate the methods and data together in a single unit. In class we...
Read more