4
The Perceptron

The Perceptron. Perceptron Pattern Classification One of the purposes that neural networks are used for is pattern classification. Once the neural network

Embed Size (px)

Citation preview

Page 1: The Perceptron. Perceptron Pattern Classification One of the purposes that neural networks are used for is pattern classification. Once the neural network

The Perceptron

Page 2: The Perceptron. Perceptron Pattern Classification One of the purposes that neural networks are used for is pattern classification. Once the neural network

Perceptron

Pattern Classification One of the purposes that neural networks are used for is pattern

classification. Once the neural network has been trained using a learning

algorithm and a training set the neural network can classify an input vector as either belonging to a particular class or category or not.

Each training case is an input vector and target value pair. The input vector is composed of binary or bipolar values. The target value is a 1 if the pattern represented by the input

vector belongs to the class and a 0 or -1 if the input vector does not belong to the class.

The perceptron learning algorithm is often used for purposes of pattern classification.

The perceptron learning algorithm works better with bipolar rather than binary values.

Page 3: The Perceptron. Perceptron Pattern Classification One of the purposes that neural networks are used for is pattern classification. Once the neural network

Perceptron

Perceptron

.

.

x1

x2

xn

y

w1

w2

wn

1

b

Page 4: The Perceptron. Perceptron Pattern Classification One of the purposes that neural networks are used for is pattern classification. Once the neural network

Perceptron

Exercises Apply the perceptron algorithm to determine the

weights for the following functions, given theta is 0 and alpha is 1: AND OR

Train a perceptron neural network to store the following patterns: (1 -1 1) and (1 1 -1) where the first pattern belongs to class and the second does not. Test the neural network on the following patterns on the following noisy patterns: (0 -1 1), (0, 1, -1). Alpha is 1 and theta is 0.