def __init__(self, dimensions, onIteration=None, initial_weights=None):
self.dimensions = dimensions
if initial_weights is not None:
self.weights = initial_weights
self.weights = normalize([0.000001 for _ in xrange(dimensions)])
self.bias = 0.0
self.cached_weights = normalize([0.000001 for _ in xrange(dimensions)])
self.cached_bias = 0.0
self.counter = 1.0
self.onIteration = onIteration
def gen_random_training_data(num_examples, dimensions, initial_weights=None, initial_bias=None):
if initial_weights is None:
weights = normalize([random.uniform(-1, 1) for _ in xrange(dimensions)])
else:
weights = initial_weights
if initial_bias is None:
bias = random.uniform(-1, 1)
else:
bias = initial_bias
training_data = []
for _ in xrange(num_examples):
feature_vector = [random.uniform(-1, 1) for _ in xrange(dimensions)]
classification = classify(feature_vector, weights, bias)
training_data.append((feature_vector, classification))
return training_data, weights, bias