How can I graph a numerical function using Python and Matplotlib?

Question

I'm trying to graph the Sigmoid Function used in machine learning by using the Matplotlib library. My problem is that I haven't visualized a mathematical function before so I'm humbly asking for your guidance.

I've tried to directly plot the following function:

def Sigmoid(x):
  a=[]
  for i in x:
    a.append(1/(1+math.exp(-i)))
  return a

using the command plt.plot(Sigmoid). But that gave me the error:

TypeError: float() argument must be a string or a number, not 'function'

The final result should look something like this:

Guys, non of these solutions worked for me. So, I added two lines of code based on the answers I received: x = np.linspace(-5, 5) y = Sigmoid(x). Apparently, I had to define an interval range for the x-axis and let y be the function itself. This means that for every value the function ouputs, Matplotlib will graph it on the interval -5<x<5... — Sobhan
– Sobhan, Commented May 8, 2019 at 6:19

PMende · Accepted Answer · 2019-05-02 19:51:55Z

4

import numpy as np
import matplotlib.pyplot as plt

def sigmoid(arr, scale=1):
    arr = np.asarray(arr)
    result = 1/(1 + np.exp(-arr*scale))
    return result

x = np.linspace(-5, 5)
y = sigmoid(x)

fig, ax = plt.subplots()
ax.plot(x, y)

Result:

The ax.plot method takes a pair of 1-D array-likes that are of the same length to create the lines. Matplotlib is not like Mathematica in which you can give an analytic function and a domain of its arguments. You have to give (in this case) x-y pairs (or rather, lists/arrays that can be turned into x-y pairs) And in this case, order matters.

answered May 2, 2019 at 19:51

PMende

5,4903 gold badges21 silver badges26 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

Sheldore Over a year ago

x is already an array. Why are you again converting it to an array inside the function ?

PMende Over a year ago

@Sheldore The input doesn't necessarily have to be an array. np.asarray will use the array as-is (i.e., won't copy it) if it's already a numpy array. Otherwise, it converts it.

Sheldore Over a year ago

My comment was in the context of your posted answer, not a generality.

gboffi · Accepted Answer · 2019-05-02 19:53:37Z

3

Sigmoid is a function, Matplotlib expects numerical values, i.e., the results of a function evaluation, e.g.

x = [i/50 - 1 for i in range(101)]
plt.plot(x, Sigmoid(x))

That said, you probably want to familiarize with the Numpy library

import matplotlib.pyplot as plt
import numpy as np

x = np.linspace(-1, 1, 101)
plt.plot(x, 1/(1+np.exp(-x))

answered May 2, 2019 at 19:53

gboffi

25.4k10 gold badges62 silver badges97 bronze badges

Collectives™ on Stack Overflow

How can I graph a numerical function using Python and Matplotlib?

2 Answers 2

3 Comments

Comments

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

3 Comments

Comments

Related