How to plot sigmoid curve based on linear regression results?

Question

I have a linear regression classifier. I want to implement and plot a sigmoid curve using the parameters (slope and inrtercept) of my linear regression model but i am getting some error

code:

import pandas as pd
import seaborn as sb
import numpy as np
import matplotlib.pyplot as plt
from sklearn.linear_model import LinearRegression
import math

glucose1 = [i for i in range(100,123)]
glucose2 = [i for i in range(130,140)]
glucose = glucose1 + glucose2
diabetes = [1 if i>126 else 0 for i in glucose]
Data = pd.DataFrame(list(zip(glucose,diabetes)), columns = ['Glucose','Diabetes'])


Data.head()

Y = Data['Diabetes'].values
X = Data['Glucose'].values

LR = LinearRegression()
LR.fit(X.reshape(-1,1),Y)


def sigmoid(b0,b1,x):
    a = []
    for item in x:
        a.append(1/(1 + math.exp(-(b0 + b1 * x))))
    return a

sig = sigmoid(LR.intercept_, LR.coef_, X)

plt.plot(X,sig)
plt.show()

the error message:

TypeError Traceback (most recent call last) in () 28 return a 29 ---> 30 sig = sigmoid(LR.intercept_, LR.coef_, X) 31 32 plt.plot(X,sig)

in sigmoid(b0, b1, x) 25 a = [] 26 for item in x: ---> 27 a.append(1/(1 + math.exp(-(b0 + b1 * x)))) 28 return a 29

TypeError: only size-1 arrays can be converted to Python scalars

Phablulo Joel Phablulo Joel · Accepted Answer · 2019-11-29T09:23:59

for item in x:
        a.append(1/(1 + math.exp(-(b0 + b1 * x))))
    return a

I think you have a typo. x is an array but you're multiplying it by b1 and addind to b0. math.exp works on scalars, not on arrays.

How to plot sigmoid curve based on linear regression results?

1 Answers