def diffusion_trend(ind):
iterations, iterations_mitigations = evaluate.evaluate_individual(
toolbox.compile(ind),
m=model,
traveler_set=travelers,
mvc_set=minimal_vertex_cover,
vert_avg_dist=vertex_average_distance,
number_vertex_shortest=number_shortest_paths,
Page_Rank=page_rank,
Cluster_Coeff=cluster_coef,
avg_degree=average_degree,
short_dist=shortest_distances,
avg_dist=average_distance,
total_iterations=ITERATIONS,
measure_every=MEASURE_EVERY,
mitigations_per_measure=MITIGATIONS_PER_MEASURE,
rollover=ROLLOVER,
use_all=USE_ALL)
trends = model.build_trends(iterations)
# Visualization
viz = DiffusionTrend(model, trends)
viz.plot()
viz = DiffusionPrevalence(model, trends)
viz.plot()
return iterations, trends
def evaulate_trained_individual(self):
"""testing evaluate.train_individual_fast"""
# create toolbox using deap_functions
toolbox = deap_functions.create_toolbox()
# loading data
# input_file = '../data/sample.csv'
input_file = '../data/train.csv'
image_size = 28
training_dataset, testing_dataset, validating_dataset, training_labels, testing_labels, validating_labels = \
data_load.data_load(input_file, image_size)
# create one individual
individual = toolbox.individual()
# run training
accuracy, protobuf = evaluate.train_individual_fast(
individual, validating_dataset, validating_labels)
# run evaluation in order to check if accuracy of the individual is equal to accuracy
# calculated as part of the training
accuracy = evaluate.evaluate_individual(individual, protobuf,
testing_dataset,
testing_labels)
print("accuracy from training: ", accuracy)
print("type of model :", type(protobuf))
import sys
print(protobuf)
self.assertTrue(hasattr(toolbox, 'individual'))
print("size of model : ", sys.getsizeof(protobuf))
print("deep size of the the model : ", deep_getsizeof(protobuf, set()))
# use model in protobuf to evaluate individual
accuracy = evaluate.evaluate_individual(individual, protobuf,
testing_dataset,
testing_labels)
print("accuracy from evaluation: ", accuracy)
page_rank = get_all_page_rank(model)
cluster_coef = clustering_coefficient(model)
# Evaluate the function
# If we are doing the non mitigation
# we must not do a secondary strategy
if FUNCTION.__name__ != "mitigation_none":
iterations, iterations_mitigations = evaluate.evaluate_individual(
FUNCTION,
m=model,
traveler_set=travelers,
mvc_set=minimal_vertex_cover,
vert_avg_dist=vertex_average_distance,
number_vertex_shortest=number_shortest_paths,
Page_Rank=page_rank,
Cluster_Coeff=cluster_coef,
avg_degree=average_degree,
short_dist=shortest_distances,
avg_dist=average_distance,
total_iterations=ITERATIONS,
measure_every=MEASURE_EVERY,
mitigations_per_measure=MITIGATIONS_PER_MEASURE,
rollover=ROLLOVER,
use_all=USE_ALL)
else:
iterations, iterations_mitigations = evaluate.evaluate_individual(
FUNCTION,
m=model,
traveler_set=travelers,
mvc_set=minimal_vertex_cover,
vert_avg_dist=vertex_average_distance,