perceptron.py

# -*- coding: utf-8 -*-
"""PERCEPTRON

Automatically generated by Colab.

Original file is located at
    https://colab.research.google.com/drive/1Usd8JvlUB4HVYvSS3TRwkf_9M9Cr-doC

# ANN - Artificial Neural Network

Usar a arquitetura perceptron para resolver o problema de portas lógicas
"""

from sklearn.linear_model import Perceptron
import numpy as np

"""## Logic gate: AND"""

data = [[0, 0], [0, 1], [1, 0], [1, 1]]
target = [0, 0, 0, 1]


classifier = Perceptron(max_iter=50000)
classifier.fit(data, target)
print('Accuracy: ', classifier.score(data, target))

predictions = classifier.predict(data)
print('Predictions: ', predictions)

"""## Logic gate: OR"""

data = [[0, 0], [0, 1], [1, 0], [1, 1]]
target = [0, 1, 1, 1]


classifier = Perceptron(max_iter=50000)
classifier.fit(data, target)
print('Accuracy: ', classifier.score(data, target))

predictions = classifier.predict(data)
print('Predictions: ', predictions)

"""## Logic gate: XOR"""

# WITH A SINGLE PERCEPTRON

data = [[0, 0], [0, 1], [1, 0], [1, 1]]
target = [0, 1, 1, 0]


classifier = Perceptron(max_iter=999000000) # 999 million
classifier.fit(data, target)
print('Accuracy: ', classifier.score(data, target))

predictions = classifier.predict(data)
print('Predictions: ', predictions)

# WITH TWO PERCEPTRONS

data = [[0, 0], [0, 1], [1, 0], [1, 1]]
targetAND = [0, 0, 0, 1]
targetOR = [0, 1, 1, 1]
targetXOR = [0, 1, 1, 0]

# aprender AND
classifierAND = Perceptron(max_iter=500000)
classifierAND.fit(data, targetAND)
print('Accuracy AND: ', classifierAND.score(data, targetAND))

# aprender OR
classifierOR = Perceptron(max_iter=500000)
classifierOR.fit(data, targetOR)
print('Accuracy OR: ', classifierOR.score(data, targetOR))

first_layer_output = np.column_stack([
    classifierOR.predict(data),   # OR output
    classifierAND.predict(data)   # AND output
])

classifierXOR = Perceptron(max_iter=999000000)
classifierXOR.fit(first_layer_output, targetXOR)

# Predict XOR
xor_predictions = classifierXOR.predict(first_layer_output)
print('Accuracy XOR: ', classifierXOR.score(first_layer_output, targetXOR))
print("XOR Predictions:", xor_predictions)

"""## Logic gate: XOR with MLP"""

from sklearn.neural_network import MLPClassifier

data = [[0, 0], [0, 1], [1, 0], [1, 1]]
target = [0, 1, 1, 0]

# Create an MLP classifier with one hidden layer of 2 neurons
mlp_classifier = MLPClassifier(hidden_layer_sizes=(2,), activation='logistic', solver='lbfgs', max_iter=10000)

# Train the classifier
mlp_classifier.fit(data, target)

# Evaluate the classifier
print('Accuracy: ', mlp_classifier.score(data, target))

# Make predictions
predictions = mlp_classifier.predict(data)
print('Predictions: ', predictions)