def TDUpdate(state, nextState, reward, w, eta=1e-1):
features = util.extractFeatures(state)
value = util.logisticValue(w, features)
residual = reward - value
if nextState is not None:
nextFeatures = util.extractFeatures(nextState)
residual += util.logisticValue(w, nextFeatures)
gradient = value * (1 - value) * features
newWeights = w + eta * residual * gradient
return newWeights
def TDUpdate(state, nextState, reward, w, eta=1e-1):
features = util.extractFeatures(state)
value = util.logisticValue(w, features)
residual = reward - value
if nextState is not None:
nextFeatures = util.extractFeatures(nextState)
residual += util.logisticValue(w, nextFeatures)
gradient = value * (1 - value) * features
newWeights = w + eta * residual * gradient
return newWeights
def getValueRec(self, agent, game, card, depth, alpha, beta):
otherAgent = int(not agent)
gameClone = copy.deepcopy(game)
gameClone.playCard(otherAgent, card)
if gameClone.roundOver():
gameClone.endRound()
if gameClone.gameOver() and gameClone.isWinner(self.playerNum):
return 1
elif gameClone.gameOver() and gameClone.isLoser(self.playerNum):
return 0
elif depth == 0:
state = gameClone.getState(agent)
features = util.extractFeatures(state)
if agent == gameClone.attacker:
weights = self.w_atk
else:
weights = self.w_def
return util.logisticValue(weights, features)
if otherAgent == self.playerNum:
depth -= 1
if agent == gameClone.attacker:
cards = gameClone.getAttackOptions(agent)
else:
cards = gameClone.getDefendOptions(agent)
if agent == self.playerNum:
v = float('-inf')
for card in cards:
v = max(
v,
self.getValueRec(otherAgent, gameClone, card, depth, alpha,
beta))
if v >= beta:
return v
alpha = max(alpha, v)
return v
else:
v = float('+inf')
for card in cards:
v = min(
v,
self.getValueRec(otherAgent, gameClone, card, depth, alpha,
beta))
if v <= alpha:
return v
beta = min(beta, v)
return v
def getValue(self, card, game):
gameClone = copy.deepcopy(game)
gameClone.playCard(self.playerNum, card)
state = gameClone.getState(self.playerNum)
if state['isAttacker']:
if card == dk.Durak.END_ROUND:
state['isAttacker'] = False
weights = self.w_def
else:
weights = self.w_atk
else:
weights = self.w_def
state['hand'].addCards(state['table'].getCards())
features = util.extractFeatures(state)
return util.logisticValue(weights, features)
def getValue(self, card, game):
gameClone = copy.deepcopy(game)
gameClone.playCard(self.playerNum, card)
state = gameClone.getState(self.playerNum)
if state['isAttacker']:
if card == dk.Durak.END_ROUND:
state['isAttacker'] = False
weights = self.w_def
else:
weights = self.w_atk
else:
weights = self.w_def
state['hand'].addCards(state['table'].getCards())
features = util.extractFeatures(state)
return util.logisticValue(weights, features)
def getValueRec(self, agent, game, card, depth, alpha, beta):
otherAgent = int(not agent)
gameClone = copy.deepcopy(game)
gameClone.playCard(otherAgent, card)
if gameClone.roundOver():
gameClone.endRound()
if gameClone.gameOver() and gameClone.isWinner(self.playerNum):
return 1
elif gameClone.gameOver() and gameClone.isLoser(self.playerNum):
return 0
elif depth == 0:
state = gameClone.getState(agent)
features = util.extractFeatures(state)
if agent == gameClone.attacker:
weights = self.w_atk
else:
weights = self.w_def
return util.logisticValue(weights, features)
if otherAgent == self.playerNum:
depth -= 1
if agent == gameClone.attacker:
cards = gameClone.getAttackOptions(agent)
else:
cards = gameClone.getDefendOptions(agent)
if agent == self.playerNum:
v = float('-inf')
for card in cards:
v = max(v, self.getValueRec(otherAgent, gameClone, card, depth, alpha, beta))
if v >= beta:
return v
alpha = max(alpha, v)
return v
else:
v = float('+inf')
for card in cards:
v = min(v, self.getValueRec(otherAgent, gameClone, card, depth, alpha, beta))
if v <= alpha:
return v
beta = min(beta, v)
return v
import pandas as pd
import util
from sklearn.cross_validation import train_test_split
from sklearn.ensemble import GradientBoostingRegressor,RandomForestRegressor
from sklearn.externals import joblib
from sklearn.metrics import classification_report
#build 7's model after case study
data7=pd.read_excel('201707_label new.xlsx')
#data7=data7[data7['是否接通']==1]
#data7=data7[['存折计划' in c for c in data7['租机计划']]]
#data7['跪舔']=data7['跪舔']-data7['label']#这个feature使f1-score增加0.02
data7.loc[data7['跪舔']>=1,'跪舔']=1
data7=util.preprocessing(data7)
X=features=util.extractFeatures(data7)
y=data7['label']
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42)
model=GradientBoostingRegressor()
model.fit(X_train, y_train)
print(model.score(X_train,y_train))
print(model.score(X_test,y_test))
y_pre=model.predict(X_test)
a=pd.DataFrame({"pre":y_pre,"y_test":y_test}).merge(data7,how='left',left_index=True,right_index=True)[['pre','跪舔','y_test']]
for k in range(50,105,5):
k=k/100
a['pre_weight_{}'.format(k)]=k*a.pre+(1-k)*a['跪舔']
util.showReport(a['pre_weight_{}'.format(k)],a.y_test,30)
#util.showFigure1(y_pre,y_test)
#util.showPRfigure(y_pre,y_test)
#util.showReport(y_pre,y_test,20)
'''
y_pre=model.predict(X_test)
util.showFigure1(y_pre,y_test)
util.showPRfigure(y_pre,y_test)
util.showReport(y_pre,y_test,50)
'''
#predict 9 with 8's model
data8 = pd.read_excel('8月租机到期数据-结果.xlsx')
data8 = data8[data8['是否接通'] == 1]
data8['跪舔'] = data8['跪舔'] - data8['label']
data8.loc[data8['跪舔'] >= 1, '跪舔'] = 1
data9 = pd.read_csv('201709-租机到期数据.csv', encoding='gbk')
data9.loc[data9['跪舔'] >= 1, '跪舔'] = 1
data = pd.concat((data8, data9), axis=0, join='inner', ignore_index=True)
data = util.preprocessing(data)
X = features = util.extractFeatures(data, num_train=data8.shape[0])
X_train = X[:data8.shape[0]]
X_test = X[data8.shape[0]:]
y_train = data8['label']
model = GradientBoostingRegressor()
model.fit(X_train, y_train)
print(model.score(X_train, y_train))
y_pre = model.predict(X_test)
data9['label'] = y_pre
data9['label'] = 0.9 * data9['label'] + 0.1 * data9['跪舔']
data9.sort_values(by='label', ascending=False, inplace=True)
data9.to_csv('201709_score.csv')
#slice
data9 = data9[[
'客户等级', '客户名称', '客户证件类型', '用户状态', '服务号码', '用户在网时长（月）', '入网时间', '年龄',
def generateDataAndLabels(inputFiles,
matrixOutputFile,
controlMatrixOutputFile,
labelOutputFile,
dateRange=None):
if os.path.isfile(matrixOutputFile) and os.path.isfile(
controlMatrixOutputFile) and os.path.isfile(labelOutputFile):
print('Opening old file')
with open(matrixOutputFile, 'rb') as f:
pubRelations = pickle.load(f)
with open(controlMatrixOutputFile, 'rb') as f:
controlPubRelations = pickle.load(f)
with open(labelOutputFile, 'rb') as f:
labels = pickle.load(f)
return pubRelations, controlPubRelations, labels
if os.path.isfile('temp' + labelOutputFile):
with open('temp' + labelOutputFile, 'rb') as f:
publicationList = pickle.load(f)
else:
publicationList = PubmedStringParser.getAllData(inputFiles)
for i in range(len(publicationList)):
publicationList[i].genRetrospectiveCitations()
publicationList[i].getCitedByPMCitations()
with concurrent.futures.ThreadPoolExecutor(
max_workers=constants.MAX_WORKERS) as executor:
future_to_article = {
executor.submit(publicationList[i].genCitedBy):
publicationList[i].title
for i in range(len(publicationList))
}
for future in concurrent.futures.as_completed(future_to_article):
articleTitle = future_to_article[future]
try:
data = future.result()
except Exception as exc:
print('%r generated an exception: %s' %
(articleTitle, exc))
else:
print('Graph retrieved for %s' % articleTitle)
if data is None:
print('But the data is none')
else:
print(data)
with open('temp' + labelOutputFile, 'wb') as f:
pickle.dump(publicationList, f)
# for publication in publicationList:
# publication.pubPrint()
# Experimental vs control publication relations (eg TF-IDF vs count)
publicationRelationships = []
controlPubRelations = []
labels = []
for i in range(len(publicationList)):
if len(publicationList[i].abstract) == 0: continue
# if len(publicationList[i].citedBy) == 0: continue
for j in range(i + 1, len(publicationList)):
if len(publicationList[j].abstract) == 0: continue
# if len(publicationList[j].citedBy) == 0: continue
if dateRange == None or util.dateDifference(
publicationList[i].date,
publicationList[j].date) <= dateRange:
publicationRelationships.append(
util.extractFeatures(publicationList[i],
publicationList[j], 'tfidf'))
controlPubRelations.append(
util.extractFeatures(publicationList[i],
publicationList[j], 'count'))
citedByI = publicationList[i].citedBy
citedByJ = publicationList[j].citedBy
pmCitedByI = publicationList[i].PMCitedBy
pmCitedByJ = publicationList[j].PMCitedBy
if len(citedByI & citedByJ) == 0 and len(pmCitedByI
& pmCitedByJ) == 0:
labels.append(0)
else:
labels.append(1)
# Convert to np array
publicationRelationships = np.array(publicationRelationships)
controlPubRelations = np.array(controlPubRelations)
labels = np.array(labels)
print('Saving new file')
with open(matrixOutputFile, 'wb') as f:
pickle.dump(publicationRelationships, f)
with open(controlMatrixOutputFile, 'wb') as f:
pickle.dump(controlPubRelations, f)
with open(labelOutputFile, 'wb') as f:
pickle.dump(labels, f)
return publicationRelationships, controlPubRelations, labels
