def strNumClassRules(rules):
list_string = []
for consequent in mlem2.getEstimatedClass(rules):
rules_consequent = mlem2.getRulesClass(rules, consequent)
num = len(rules_consequent)
list_string.append(str(num))
return (",".join(list_string))
def strNumClassRules(rules):
list_string = []
for consequent in mlem2.getEstimatedClass(rules):
rules_consequent = mlem2.getRulesClass(rules, consequent)
num = len(rules_consequent)
list_string.append(str(num))
return ",".join(list_string)
def getAccurayRecall(rules, decision_table, list_judgeNominal):
result = []
consequents = mlem2.getEstimatedClass(rules)
for consequent in consequents :
target_rules = mlem2.getRulesClass(rules, consequent)
target_objects = []
for rule in target_rules :
match_objects = decision_table.apply(lambda obj: isExplainRule(obj, rule, list_judgeNominal), axis=1)
index_objects = decision_table[match_objects].index.tolist()
target_objects.extend(index_objects)
target_objects = list(set(target_objects))
estimated_classes = decision_table[decision_table.columns[-1]].ix[target_objects]
accuracy = sum(estimated_classes == consequent)/ len(target_objects)
recall = sum(estimated_classes == consequent) / sum(decision_table[decision_table.columns[-1]] == consequent)
result.append((accuracy, recall))
return(result)
def getAccurayRecall(rules, decision_table, list_judgeNominal):
result = []
consequents = mlem2.getEstimatedClass(rules)
for consequent in consequents:
target_rules = mlem2.getRulesClass(rules, consequent)
target_objects = []
for rule in target_rules:
match_objects = decision_table.apply(
lambda obj: isExplainRule(obj, rule, list_judgeNominal),
axis=1)
index_objects = decision_table[match_objects].index.tolist()
target_objects.extend(index_objects)
target_objects = list(set(target_objects))
estimated_classes = decision_table[
decision_table.columns[-1]].ix[target_objects]
accuracy = sum(estimated_classes == consequent) / len(target_objects)
recall = sum(estimated_classes == consequent) / sum(
decision_table[decision_table.columns[-1]] == consequent)
result.append((accuracy, recall))
return (result)
def MLEM2_delEAlphaRule_LERS(FILENAME, iter1, iter2, DELFUN, CLASS,
ATTRIBUTE_VALUE, alpha):
print(datetime.now().strftime('%Y/%m/%d %H:%M:%S') + ' ' + FILENAME + ' ' +
str(iter1) + ' ' + str(iter2) + ' ' + DELFUN.__name__ + ' ' +
strAttributeValue(ATTRIBUTE_VALUE) + ' ' + str(alpha) + ' ' +
"START")
# rule induction and rule save
fullpath_filename = DIR_UCI + '/' + FILENAME + '/rules/' + 'rules_' + str(
iter1) + '-' + str(iter2) + '.pkl'
rules = mlem2.loadPickleRules(fullpath_filename) if os.path.isfile(
fullpath_filename) else mlem2.getRulesByMLEM2(FILENAME, iter1, iter2)
if not os.path.isfile(fullpath_filename):
mlem2.savePickleRules(rules, fullpath_filename)
# train data setup
decision_table_train, decision_class = getData(FILENAME,
iter1,
iter2,
T="train")
list_judgeNominal = getJudgeNominal(decision_table_train, FILENAME)
# alpha差別的なルールの基本条件削除 or ルールを削除
if CLASS != "ALL":
rules_target = mlem2.getRulesClass(rules, CLASS)
rules_nontarget = mlem2.getRulesClass(rules, CLASS, judge=False)
for attr in ATTRIBUTE_VALUE:
for e in ATTRIBUTE_VALUE[attr]:
rules_target = DELFUN(rules_target, attr, e,
decision_table_train, list_judgeNominal,
alpha)
rules_target.extend(rules_nontarget)
rules = rules_target
else:
for attr in ATTRIBUTE_VALUE:
for e in ATTRIBUTE_VALUE[attr]:
rules = DELFUN(rules, attr, e, decision_table_train,
list_judgeNominal, alpha)
print(datetime.now().strftime('%Y/%m/%d %H:%M:%S') + ' ' + FILENAME + ' ' +
str(iter1) + ' ' + str(iter2) + ' ' + DELFUN.__name__ + ' ' +
strAttributeValue(ATTRIBUTE_VALUE) + ' ' + str(alpha) + ' ' +
"RULES")
# test data setup
decision_table_test, decision_class = getData(FILENAME,
iter1,
iter2,
T="test")
list_judgeNominal = getJudgeNominal(decision_table_test, FILENAME)
# predict by LERS
predictions = LERS.predictByLERS(rules, decision_table_test,
list_judgeNominal)
# 正答率を求める
accuracy = accuracy_score(decision_class, predictions)
# rules の数を求める
num = len(rules)
# 各クラスのrulesの数を求める
num_class = strNumClassRules(rules)
# 平均の長さを求める
mean_length = mlem2.getMeanLength(rules)
# 平均支持度と平均確信度を求める
list_judgeNominal = getJudgeNominal(decision_table_train, FILENAME)
mean_support, mean_conf = LERS.getSupportConfidenceRules(
rules, decision_table_train, list_judgeNominal)
# AccとRecallを求める
acc_recall = LERS.getAccurayRecall(rules, decision_table_train,
list_judgeNominal)
# ファイルにsave
savepath = DIR_UCI + '/' + FILENAME + '/fairness/02_alpha_preserve/MLEM2_delEAlphaRule_LERS.csv'
with open(savepath, "a") as f:
f.writelines(
'MLEM2_delEAlphaRule_LERS,{DELFUN},{CLASS},{FILENAME},{ATTRIBUTE_VALUE},{alpha},{iter1},{iter2},{acc},{num},{num_class},{mean_length},{mean_support},{mean_conf},{acc_recall}'
.format(DELFUN=DELFUN.__name__,
CLASS=CLASS,
FILENAME=FILENAME,
ATTRIBUTE_VALUE=strAttributeValue(ATTRIBUTE_VALUE),
alpha=alpha,
iter1=iter1,
iter2=iter2,
acc=accuracy,
num=num,
num_class=num_class,
mean_length=mean_length,
mean_support=mean_support,
mean_conf=mean_conf,
acc_recall=strAccRecall(rules, acc_recall)) + "\n")
print(datetime.now().strftime('%Y/%m/%d %H:%M:%S') + ' ' + FILENAME + ' ' +
str(iter1) + ' ' + str(iter2) + ' ' + DELFUN.__name__ + ' ' +
strAttributeValue(ATTRIBUTE_VALUE) + ' ' + str(alpha) + ' ' + "END")
return (0)
def MLEM2_delEAlphaRule_LERS(FILENAME, iter1, iter2, DELFUN, CLASS, ATTRIBUTE_VALUE, alpha):
print(
datetime.now().strftime("%Y/%m/%d %H:%M:%S")
+ " "
+ FILENAME
+ " "
+ str(iter1)
+ " "
+ str(iter2)
+ " "
+ DELFUN.__name__
+ " "
+ strAttributeValue(ATTRIBUTE_VALUE)
+ " "
+ str(alpha)
+ " "
+ "START"
)
# rule induction and rule save
fullpath_filename = DIR_UCI + "/" + FILENAME + "/rules/" + "rules_" + str(iter1) + "-" + str(iter2) + ".pkl"
rules = (
mlem2.loadPickleRules(fullpath_filename)
if os.path.isfile(fullpath_filename)
else mlem2.getRulesByMLEM2(FILENAME, iter1, iter2)
)
if not os.path.isfile(fullpath_filename):
mlem2.savePickleRules(rules, fullpath_filename)
# train data setup
decision_table_train, decision_class = getData(FILENAME, iter1, iter2, T="train")
list_judgeNominal = getJudgeNominal(decision_table_train, FILENAME)
# alpha差別的なルールの基本条件削除 or ルールを削除
if CLASS != "ALL":
rules_target = mlem2.getRulesClass(rules, CLASS)
rules_nontarget = mlem2.getRulesClass(rules, CLASS, judge=False)
for attr in ATTRIBUTE_VALUE:
for e in ATTRIBUTE_VALUE[attr]:
rules_target = DELFUN(rules_target, attr, e, decision_table_train, list_judgeNominal, alpha)
rules_target.extend(rules_nontarget)
rules = rules_target
else:
for attr in ATTRIBUTE_VALUE:
for e in ATTRIBUTE_VALUE[attr]:
rules = DELFUN(rules, attr, e, decision_table_train, list_judgeNominal, alpha)
print(
datetime.now().strftime("%Y/%m/%d %H:%M:%S")
+ " "
+ FILENAME
+ " "
+ str(iter1)
+ " "
+ str(iter2)
+ " "
+ DELFUN.__name__
+ " "
+ strAttributeValue(ATTRIBUTE_VALUE)
+ " "
+ str(alpha)
+ " "
+ "RULES"
)
# test data setup
decision_table_test, decision_class = getData(FILENAME, iter1, iter2, T="test")
list_judgeNominal = getJudgeNominal(decision_table_test, FILENAME)
# predict by LERS
predictions = LERS.predictByLERS(rules, decision_table_test, list_judgeNominal)
# 正答率を求める
accuracy = accuracy_score(decision_class, predictions)
# rules の数を求める
num = len(rules)
# 各クラスのrulesの数を求める
num_class = strNumClassRules(rules)
# 平均の長さを求める
mean_length = mlem2.getMeanLength(rules)
# 平均支持度と平均確信度を求める
list_judgeNominal = getJudgeNominal(decision_table_train, FILENAME)
mean_support, mean_conf = LERS.getSupportConfidenceRules(rules, decision_table_train, list_judgeNominal)
# AccとRecallを求める
acc_recall = LERS.getAccurayRecall(rules, decision_table_train, list_judgeNominal)
# ファイルにsave
savepath = DIR_UCI + "/" + FILENAME + "/fairness/02_alpha_preserve/MLEM2_delEAlphaRule_LERS.csv"
with open(savepath, "a") as f:
f.writelines(
"MLEM2_delEAlphaRule_LERS,{DELFUN},{CLASS},{FILENAME},{ATTRIBUTE_VALUE},{alpha},{iter1},{iter2},{acc},{num},{num_class},{mean_length},{mean_support},{mean_conf},{acc_recall}".format(
DELFUN=DELFUN.__name__,
CLASS=CLASS,
FILENAME=FILENAME,
ATTRIBUTE_VALUE=strAttributeValue(ATTRIBUTE_VALUE),
alpha=alpha,
iter1=iter1,
iter2=iter2,
acc=accuracy,
num=num,
num_class=num_class,
mean_length=mean_length,
mean_support=mean_support,
mean_conf=mean_conf,
acc_recall=strAccRecall(rules, acc_recall),
)
+ "\n"
)
print(
datetime.now().strftime("%Y/%m/%d %H:%M:%S")
+ " "
+ FILENAME
+ " "
+ str(iter1)
+ " "
+ str(iter2)
+ " "
+ DELFUN.__name__
+ " "
+ strAttributeValue(ATTRIBUTE_VALUE)
+ " "
+ str(alpha)
+ " "
+ "END"
)
return 0