def evaluate_circuit(n = 208, eval_perceptron = True):
test_idxs = np.random.choice(data.shape[0], n, replace = False)
correct_circuit = 0
accuracy_circuit = 0.0
correct_perceptron = 0
accuracy_perceptron = 0.0
#test_idxs = 208
for i in test_idxs:
#print(trial(weights, p_weights, ap_weights, i, plot = False))
#print(data[i][-1])
#print()
trial_result = trial(weights, p_weights, ap_weights, i, plot = False, num = 1)
if trial_result[0] == data[i][-1]:
correct_circuit += 1
if(eval_perceptron):
#This uses multiple noisy training rows over n epochs, should we instead train the perceptron on a single noisy row n times?
perceptron_training_data = data[test_idxs, :-1]
perceptron_noisy_data = perceptron_training_data + np.random.rand(n, 60)
classes = data[[test_idxs], [-1]]
classes = classes.reshape(n, 1)
perceptron_noisy_data = np.append(perceptron_noisy_data, classes, axis = 1)
perceptron = p.gradient_descent(perceptron_noisy_data, 0.1, trial_result[1])
if p.predict(data[i], perceptron) == data[i][-1]:
correct_perceptron += 1
#Train perceptron on noisy data for as many epochs as it took the circuit to decide
#determine perceptrons accuracy
#print(perceptron)
accuracy_circuit = (correct_circuit / n) * 100
accuracy_perceptron = (correct_perceptron / n) * 100
print("circuit accuracy: ", accuracy_circuit)
print("perceptron accuracy:", accuracy_perceptron)
if (running_accuracy < accuracy_threshold):
left_off_index = i
break
context_retrain = True
retrain_weights = np.empty((100,3))
#cmap = colors.LinearSegmentedColormap('custom', cdict)
for i in range (left_off_index, test.shape[0]):
context_2_timer += 1
datum = data[i]
if context_retrain:
#retrain a perceptron and antiperceptron
p_weights = p.gradient_descent(data[left_off_index : left_off_index + 50], 0.1, 400)
ap_weights = p.gradient_descent(data[left_off_index : left_off_index + 50], 0.1, 400, antiperceptron = True)
circuit_steps = np.ones((1, 200))
for j in range (0, 200):
#circuit_values = diffusion_predict(p_weights, ap_weights, data2[i+j,:2], data[i+j, 3])
#context_2_weights = p.gradient_descent(data[left_off_index : left_off_index + 50], 0.1, 200)
#circuit_steps[0][j] = circuit_values[1]
"""
If we use circuit_steps as
"""
circuit_steps.fill(0.2)
context_2_weights = gradient_descent_variable_eta(data2, circuit_steps, 400, context_1_weights, plot = True)
print(context_2_weights)
context_retrain = False
"""
train_data_0 = data[np.random.choice(97, 90, replace=False), :] train_data_1 = data[np.random.choice(np.arange(97, 208), 90, replace=False), :] train_data = np.concatenate((train_data_0, train_data_1), axis=0) np.random.shuffle(train_data) #def main(): #rng = np.random.default_rng() #train_data = rng.shuffle(data) #train_data2 = data[np.random.choice(data.shape[0], 154, replace = False), :] #Train perceptron, store weights #p_weights = p.gradient_descent(train_data, 0.005, 900) ap_weights = p.gradient_descent(train_data, 0.005, 900, antiperceptron=True) #p.gradient_descent(train_data, 0.005, 900, antiperceptron = True) print(ap_weights) #print(perceptron_accuracy(ap_weights, ap = True)) #print(p_weights) #print(perceptron_accuracy(p_weights)) #print(ap_weights) #return(ap_weights) #main() ###89.423% accuracy for perceptron """ [ 0.415 0.3567415 0.2803695 -0.3303615 0.1614965 0.280231 -0.0106485 -0.1240805 -0.346292 0.288558 -0.0213655 0.196917
def perceptronval():
for i in range(0,10):
pp = p.gradient_descent(train_data, (0.2 - (0.15 * i)), 700)
print(perceptron_accuracy(pp))
print(pp)
dataset = np.asarray(load_data("sonar1.all-data"))
ROW_LENGTH = len(dataset[0])
COLUMNS = len(dataset)
#noisy_data = p.noisyData(dataset)
#perceptron = p.perceptron
#antiperceptron = p.antiperceptron
p_random_indices = np.random.choice(110, size=110, replace=False)
p_train_data = dataset[p_random_indices, :110]
ap_random_indices = np.random.choice(110, size=110, replace=False)
ap_train_data = dataset[ap_random_indices, :110]
#test_data = np.delete(noisy_data, random_indices, 0)
p_weights = p.gradient_descent(p_train_data, 0.1, 200)
ap_weights = p.gradient_descent(p_train_data, 0.1, 200, antiperceptron=True)
#scores = p.evaluate_algorithm(dataset, p.perceptron, 3, .01, 500)
c_weights = np.array([
[0, -.5, 0, -.5], #1->1 2->1 3->1 4->1
[-.5, 0, -.5, 0], #1->2 2->2 3->2 4->2
[2.23, 0, 6, -6], #1->3 2->3 3->3 4->3
[0, 2.23, 0, 2.5]
]) #1->4 2->4 3->4 4->4
weights = np.array([
[0, -2, 0, -2], #1->1 2->1, 3->1, 4->1
[-2, 0, -2, 0], #1->2, 2->2, 3->2, 4->2
[1.23, 0, 2.087, 0], #1->3, 2->3, 3->3, 4->3