def main():
pop_size, gen_count, mutation = 10, 30, 0.4
#solve_methods = ["keras", "svm", "logistic", "naivebayes", \
# "randomforest", "lda"]
solve_methods = ["keras"]
accuracy_fname = "accuracy_keras.txt"
with open(accuracy_fname, "w") as fp:
for dataset in loadData():
fp.write("===================================\n")
fp.write("Filename: %s\n" % dataset['fname'])
Organism.data = dataset
Organism.count = dataset['X'].shape[1]
fp.write("Num Features: %d\n" % Organism.count)
fp.write("\n------------------------------------\n")
for solve_method in solve_methods:
print "Using solve method: ", solve_method
fp.write("Using solve method: %s\n" % solve_method)
full_accuracy = GA.full_accuracy(solve_method)
print "Accuracy using all features: ", full_accuracy.fitness
fp.write("Accuracy using all features: %f\n" % full_accuracy.fitness)
solver = GA(gen_count, pop_size, mutation, solve_method)
finalPop = solver.search()
print "Best Accuracy: ", finalPop[0].fitness
print "Subset of features used: ", finalPop[0].feature_subset
fp.write("Pop Size: %d; Generation Count: %d; Mutation Rate: %f\n" % (pop_size, gen_count, mutation))
fp.write("Best Accuracy: %f\n" % finalPop[0].fitness)
fp.write("Subset of features used: " + str(finalPop[0].feature_subset))
fp.write("\n------------------------------------\n\n")
