- What is data distribution?
- How data distribution is carried out in Numpy?

A data distribution is a listing that shows all the possible values within the given intervals or data limits. It also tells the number of occurrences of each element. Usually the distributed data is ordered from lowest to highest or alphabetically and chart or graphs are used to represent it. The probability for occurrence of any element can be calculated by applying the mathematical function on the sample space, this function is called probability density function. This function describes the probability of each element in the array and random numbers following the probability density function are called random distribution numbers.

Random module in NumPy library provides a function choice(), this function enables us to generate the elements based on a specific probability passed on as 1D array, hence there are two 1D arrays that will be passed on to the choice() function. For generating an array of a specific size and probabilities, we will have to pass choice() function three parameters. First argument will be a 1D array that will contain the elements, second parameter will be the 1D array of probabilities corresponding to each element of the 1D array being passed as first parameter while third parameter is the size of array which will be the output of the function.

To ensure correct results, we have to check that the sum of probability array equals to 1 otherwise the compiler will generate an error.

In this example, we will pass a 1D array having ten elements, a corresponding probability array and size of array being generated as output. The ten elements will having following probabilities.

1. 20 (0.2)

2. 21 (0.01)

3. 22 (0.02)

4. 23 (0.05)

5. 24 (0.1)

6. 25 (0.02)

7. 26 (0.05)

8. 27 (0.3)

9. 28 (0.1)

10. 29 (0.15)

from numpy import random as r #here we are declaring a 1D array having elements (values) sample_arr = [20,21,22,23,24,25,26,27,28,29] print('The sample array being used to produce the random distribution:', sample_arr) #here we are declaring a 1D array corresponding to each value in the element array prob_arr = [0.2,0.01,0.02,0.05,0.1,0.02,0.05,0.3,0.1,0.15] print('The probability array corresponding to sample array:', prob_arr) #passing arrays as arguments to the choice function res = r.choice(sample_arr, p=prob_arr, size=(100)) print('The result ',res)

In this example we will auto generate an array having 5 elements with value of each element between 90 and 100 and we will generate the probability array having values between 0 and 1, and pass them to the choice() function. Here you can observe that we are using size as (3,5) hence a 2D array will be produced.

from numpy import random as r #here we are auto declaring a 1D array having elements (values) sample_arr = r.randint(90,100,size=(5)) print('The 1D sample array being used to produce the random distribution:', sample_arr) #here we are declaring a 1D array corresponding to each value in the element array prob_arr = [0.1,0.5,0.08,0.3,0.02] print('The 1D probability array corresponding to sample array:', prob_arr) #passing arrays as parameters to the choice function res = r.choice(sample_arr, p=prob_arr, size=(3,5)) print('The result is a 2D array: ',res)

- 41 views

Note: Due to the size or complexity of this submission, the author has submitted it as a .zip file to shorten your download time. After downloading it, you will need a program like Winzip to decompress it.

Virus note: All files are scanned once-a-day by SourceCodester.com for viruses, but new viruses come out every day, so no prevention program can catch 100% of them.

FOR YOUR OWN SAFETY, PLEASE:

1. Re-scan downloaded files using your personal virus checker before using it.

2. NEVER, EVER run compiled files (.exe's, .ocx's, .dll's etc.)--only run source code.

## Add new comment