self.cut_off_pos = [0] * (len(self.V)+1)
for a in self.ad.number_of_v_removed:
self.cut_off_pos[a] = self.ad.number_of_v_removed[a]
if __name__ == "__main__":
increment = 1
iterations = 18
average_over = 10
iter_nums = [x*increment + increment for x in xrange(iterations)]
p_results = [0] * iterations
for j in xrange(average_over):
for i in xrange(iterations):
print j, i
a = Attack(200, 40, iter_nums[i])
a.attack(100)
p_vals = binom_stats(a.sorted_tally, 100)
p_results[i] += sum(p_vals)/len(p_vals)
p_results = [x / average_over for x in p_results]
plt.plot(iter_nums, p_results);
plt.xlabel("Threshold")
plt.ylabel("p value")
plt.show()
raw_input()
self.cut_off_pos = [0] * (len(self.V)+1)
for a in self.ad.number_of_v_removed:
self.cut_off_pos[a] = self.ad.number_of_v_removed[a]
if __name__ == "__main__":
iterations = 200
increment = 25
average_over = 10
min_p_vals = [0] * iterations
iterations_array = [i*increment + increment for i in xrange(iterations)]
for j in xrange(average_over):
a = Attack(200, 20, 3)
for i in xrange(iterations):
print j, i
a.attack(increment)
p_vals = binom_stats(a.sorted_tally, iterations_array[i])
min_p_vals[i] += min(p_vals)
min_p_vals = [x/average_over for x in min_p_vals]
fig = plt.figure()
ax = fig.add_subplot(121)
ax.plot(iterations_array, min_p_vals, 'b-', label="Minimum p-value")
ax.plot(iterations_array, [0.05] * iterations, 'r--', label="0.05 level of confidence")
ax.plot(iterations_array, [0.01] * iterations, 'm--', label="0.01 level of confidence")
ax.set_xlabel("Number of attacks")
ax.set_ylabel("Level of confidence")
ax.legend(bbox_to_anchor=(1.05, 0.5), loc=6, borderaxespad=0.)
fig.show()
