runs = 1
print('TEST r values: 5, 1, 0.5, 0.1 0.05, 0.01, 0.005, 0.001')
if (mean_std):
print('\tEach value is run ' + str(runs) +
' times. A mean and std dev for the accuracy is gathered.')
print('\nAGAINST\n\t a5a.train\n')
rng = range(runs)
r_vals = [5, 1, 0.5, 0.1, 0.05, 0.01, 0.005, 0.001]
vals = [0] * runs
for r in r_vals:
for inx in rng:
W_b = run_perceptron('res/a5a.train', r)
ep = evaluate_perceptron('res/a5a.train', W_b['W'], W_b['b'])
vals[inx] = ep['accuracy']
print('EVALUATE r =\t' + str(r))
if (mean_std):
print('Mean Accuracy:\t\t' + str(numpy.mean(vals)))
print('Std Deviation:\t\t' + str(numpy.std(vals)))
print('Updates:\t\t' + str(ep['wrong']))
print('Total Rows:\t\t' + str(ep['wrong'] + ep['right']))
print('Accuracy:\t\t' + str(ep['accuracy']))
vals = [0] * runs
print('\nAGAINST\n\t a5a.test\n')
for r in r_vals:
for inx in rng:
from Evaluate_Perceptron import evaluate_perceptron
from PerceptronAlgorithmMargin import run_perceptron_margin
r = 1
mu = 5
epoch = 10
W_b = run_perceptron_margin('res/AHU27.csv', r, mu, epoch)
ep = evaluate_perceptron('res/AHU27.csv', W_b['W'], W_b['b'])
print(W_b['W'])
print(ep)