Scribd
Upload
22 views3 pages

Ex-1 Understanding The Perceptrons

Uploaded by

pavannaidu1424
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
22 views3 pages

Ex-1 Understanding The Perceptrons

Uploaded by

pavannaidu1424
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 3

Ex:1 Understanding the Perceptron

Aim: Write a program to understand the perceptron using python.

Algorithm:
1. Import the libraries and dataset
2. The first two classes (Iris-Setosa and Iris-Versicolour) are linear separable
3. plot the data for two of the four variables
4. validate our results, we split the data into training and validation sets
5. initialize the weights and the bias for the perceptron
6. define the hyperparameters
7. training our perceptron with a for loop
8.Plot training loss and validation accuracy
Program:
# Import the libraries and dataset
import numpy as np
from sklearn.model_selection import train_test_split
import matplotlib.pyplot as plt
# We will be using the Iris Plants Database
from sklearn.datasets import load_iris
SEED = 2017
# The first two classes (Iris-Setosa and Iris-Versicolour) are linear separable
iris = load_iris()
idxs = np.where(iris.target<2)
X = iris.data[idxs]
y = iris.target[idxs]
# Let's plot the data for two of the four variables
plt.scatter(X[y==0][:,0],X[y==0][:,2], color='green', label='Iris-Setosa')
plt.scatter(X[y==1][:,0],X[y==1][:,2], color='red', label='Iris-Versicolour')
plt.title('Iris Plants Database')
plt.xlabel('sepal length in cm')
plt.ylabel('sepal width in cm')
plt.legend()
plt.show()
# To validate our results, we split the data into training and validation sets
X_train, X_val, y_train, y_val = train_test_split(X, y, test_size=0.2, random_state=SEED)
# Next, we initialize the weights and the bias for the perceptron
weights = np.random.normal(size=X_train.shape[1])
bias = 1
# Before training, we need to define the hyperparameters
learning_rate = 0.1
n_epochs = 15
np.zeros(weights.shape)
# Now, we can start training our perceptron with a for loop
del_w = np.zeros(weights.shape)
hist_loss = []
hist_accuracy = []
for i in range(n_epochs):
# We apply a simple step function, if the output is > 0.5 we predict 1, else 0
output = np.where((X_train.dot(weights)+bias)>0.5, 1, 0)
print(output)
# Compute MSE
error = np.mean((y_train-output)**2)
print("Error: ", error)
# Update weights and bias
weights-= learning_rate * np.dot((output-y_train), X_train)
bias += learning_rate * np.sum(np.dot((output-y_train), X_train))
print("Weights:", weights)
print("bias:", bias)
# Calculate MSE
loss = np.mean((output - y_train) ** 2)
hist_loss.append(loss)
# Determine validation accuracy
output_val = np.where(X_val.dot(weights)>0.5, 1, 0)
accuracy = np.mean(np.where(y_val==output_val, 1, 0))
hist_accuracy.append(accuracy)
# We've saved the training loss and validation accuracy so that we can plot them
fig = plt.figure(figsize=(8, 4))
a = fig.add_subplot(1,2,1)
imgplot = plt.plot(hist_loss)
plt.xlabel('epochs')
a.set_title('Training loss')
a=fig.add_subplot(1,2,2)
imgplot = plt.plot(hist_accuracy)
plt.xlabel('epochs')
a.set_title('Validation Accuracy')
plt.show()

Output:

Result: Hence, implementing the program to understanding the perceptron is done successfully.

You might also like