Introduction to Pattern Recognition: A Matlab Approach

2.2 The Perceptron Algorithm 33

Table 2.1 Number of Iterations Performed by the Perceptron

Algorithm in Example 2.2.1

X

1

X

2

X

3

X

4

rho =0.01 134 134 5441 No convergence

rho =0.05 5 5 252 No convergence

Step 2. To run the perceptron algorithm for X

1

with learning parameter 0.01, type

rho=0.01; % Learning rate

w_ini=[1 1 -0.5]';

[w,iter,mis_clas]=perce(X1,y1,w_ini,rho)

By altering the previous code, the perceptron algorithm may be performed on the remaining data

sets using different learning parameter values. Table 2.1 contains the results obtained by performing

the previous experiments.

Step 3. Working as in step 2, compute the w’s using the perceptron algorithm ...