def calculate(self, learner):
'''
calculate the improvements for one learner
'''
res = []
oldresult = result(self.old_path).read()
newresult = result(self.new_path).read()
for old, new in zip(oldresult[learner], newresult[learner]):
try:
res.append(new - old)
except:
continue
return res
def calculate(self, learner):
'''
calculate the improvements for one learner
'''
res = []
oldresult = result(self.old_path).read()
newresult = result(self.new_path).read()
for old, new in zip(oldresult[learner], newresult[learner]):
try:
res.append(new - old)
except:
continue
return res
def csv(self, name):
naive_learner = ['Naive_Where', 'Naive_CART', 'Naive_RF']
tuned_learner = ['Tuned_Where', 'Tuned_CART', 'Tuned_RF']
oldresult = result(self.old_path).read()
newresult = result(self.new_path).read()
res = "Name," + ",".join(newresult["Dataset"]) + '\n'
for j, aresult in enumerate([oldresult, newresult]):
for one in naive_learner:
if j == 0:
res += "Old_" + str(one) + "," + ",".join([str(i) for i in aresult[one]]) + '\n'
else:
res += "New_" + str(one) + "," + ",".join([str(i) for i in aresult[one]]) + '\n'
f = open("RX4_" + name + '.csv', "w")
f.write(res)
def tune_over_naive(path):
'''
:param path: the path of experiment results, e.g: "allf"
:return: a list of results for each learner with respect to Naive, Tuned, Grid.
'''
result_dict = result(path).read()
# pdb.set_trace()
out = []
# if "f" in path:
# result_dict["Grid_CART"] = [0,52,11,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
# result_dict["Grid_RF"] = [0,0,11,33,0,46,18,36,0,0,0,0,0,0,56,0,0]
# else:
# result_dict["Grid_CART"] = [85,85,19,93,0,0,100,0,67,60,0,60,0,0,66,0,0]
# result_dict["Grid_RF"] = [59,40,0,0,0,0,0,0,46,0,100,100,100,75,0,0,25]
# for learner in ["CART", "RF"]:
for learner in ["CART","RF"]:
res = {}
naive = None
for task in ["Naive_", "Tuned_", "Cluster_","Nbrs_","NaiveTest_"]:
name = task + learner
if task == "Naive_":
naive = name
res[name] = np.array(result_dict[name])
sorted_index = np.argsort(res[naive])
res_sorted = {}
for key, val in res.iteritems():
res_sorted[key] = val[sorted_index]
# print(str(res))
out.append(res_sorted)
return out
def csv(self, name):
naive_learner = ['Naive_Where', 'Naive_CART', 'Naive_RF']
tuned_learner = ['Tuned_Where', 'Tuned_CART', 'Tuned_RF']
oldresult = result(self.old_path).read()
newresult = result(self.new_path).read()
res = "Name," + ",".join(newresult["Dataset"]) + '\n'
for j, aresult in enumerate([oldresult, newresult]):
for one in naive_learner:
if j == 0:
res += "Old_" + str(one) + "," + ",".join(
[str(i) for i in aresult[one]]) + '\n'
else:
res += "New_" + str(one) + "," + ",".join(
[str(i) for i in aresult[one]]) + '\n'
f = open("RX4_" + name + '.csv', "w")
f.write(res)
def csv_rx4(self, name):
naive_learner = ['Naive_Where', 'Naive_CART', 'Naive_RF']
tuned_learner = ['Tuned_Where', 'Tuned_CART', 'Tuned_RF']
# oldresult = result(self.old_path).read()
newresult = result(self.new_path).read()
res = "Name," + ",".join(newresult["Dataset"]) + '\n'
for j, aresult in enumerate([newresult]):
for naive, tuned in zip(naive_learner, tuned_learner):
res += "RX4_" + str(naive) + "," + ",".join([str(i) for i in aresult[naive]]) + '\n'
res += "RX4_" + str(tuned) + "," + ",".join([str(i) for i in aresult[tuned]]) + '\n'
f = open("RX4_" + name + '.csv', "w")
f.write(res)
def csv_rx4(self, name):
naive_learner = ['Naive_Where', 'Naive_CART', 'Naive_RF']
tuned_learner = ['Tuned_Where', 'Tuned_CART', 'Tuned_RF']
# oldresult = result(self.old_path).read()
newresult = result(self.new_path).read()
res = "Name," + ",".join(newresult["Dataset"]) + '\n'
for j, aresult in enumerate([newresult]):
for naive, tuned in zip(naive_learner, tuned_learner):
res += "RX4_" + str(naive) + "," + ",".join(
[str(i) for i in aresult[naive]]) + '\n'
res += "RX4_" + str(tuned) + "," + ",".join(
[str(i) for i in aresult[tuned]]) + '\n'
f = open("RX4_" + name + '.csv', "w")
f.write(res)
