def trainModel(trainFile, testFile):
import numpy as np
import TT_ML.data_helper.data_prepare as dataProcess
data_helper = dataProcess.DataPrepare()
x_train, y_train, x_test, y_test = data_helper.loadLibSVMFile(
trainFile, testFile)
#from sklearn.linear_model import SGDClassifier
#clf=SGDClassifier(loss='log',penalty='l2',class_weight='balanced')#modified_huber
#print('TrainModel Start!')
#clf.partial_fit(x_train,y_train,classes=np.unique(y_train))
#clf.fit(x_train,y_train)
from sklearn.neighbors import KNeighborsClassifier
clf = KNeighborsClassifier(n_neighbors=7)
#clf.fit(x_train,y_train)
#pickle.dump(clf, open('./SGD_zhuyuan.pickle', 'wb'), -1)
#print('TrainModel Finish!')
print('交叉验证开始')
from sklearn.model_selection import ShuffleSplit
from sklearn.model_selection import cross_val_score
cv = ShuffleSplit(n_splits=5, test_size=0.01, random_state=0)
accs = cross_val_score(clf, x_train, y_train, cv=cv, n_jobs=5)
print('交叉验证结果:', accs)
print('交叉验证结束')
return clf
def getPredictTopN():
import TT_ML.ModelStatis.topN as topN
modelFile = './SklearnDNN_zhuyuan.pickle'
import TT_ML.data_helper.data_prepare as dataProcess
data_helper = dataProcess.DataPrepare()
x_train, y_train, x_test, y_test = data_helper.loadLibSVMFile(
'/home/jq/jeeker/zy_322/train.libsvm',
'/home/jq/jeeker/zy_322/test.libsvm')
topN.predictTop(modelFile, x_test, y_test, len(y_test))
def trainModel_MLP(trainFile, testFile):
import TT_ML.deepLearning.sklearn_DNN as sklearnDNN
import TT_ML.data_helper.data_prepare as dataProcess
data_helper = dataProcess.DataPrepare()
x_train, y_train, x_test, y_test = data_helper.loadLibSVMFile(
trainFile, testFile)
clf = sklearnDNN.MLP_normal(x_train, y_train, x_test, y_test)
pickle.dump(clf, open('./SklearnDNN_zhuyuan.pickle', 'wb'), -1)
return clf
def trainModel(trainFile, testFile):
import TT_ML.deepLearning.keras_nn as KerasNN
import TT_ML.data_helper.data_prepare as dataProcess
data_helper = dataProcess.DataPrepare()
#加载模型
x_train, y_train, x_test, y_test = data_helper.loadLibSVMFile(
trainFile, testFile)
clf = KerasNN.Kmodel_1(x_train, y_train, x_test, y_test)
return clf
def trainModel(trainFile, testFile):
import numpy as np
import TT_ML.data_helper.data_prepare as dataProcess
data_helper = dataProcess.DataPrepare()
x_train, y_train, x_test, y_test = data_helper.loadLibSVMFile(
trainFile, testFile)
from sklearn.neighbors import KNeighborsClassifier
clf = KNeighborsClassifier(n_neighbors=7)
clf.fit(x_train, y_train)
print('TrainModel Start!')
pickle.dump(clf, open(MODELPATH + '/KNN_zhuyuan.pickle', 'wb'), -1)
print('TrainModel Finish!')
return clf
def trainModel(trainFile, testFile):
import numpy as np
import TT_ML.data_helper.data_prepare as dataProcess
data_helper = dataProcess.DataPrepare()
x_train, y_train, x_test, y_test = data_helper.loadLibSVMFile(
trainFile, testFile)
from sklearn.naive_bayes import MultinomialNB
clf = MultinomialNB()
clf.partial_fit(x_train, y_train, classes=np.unique(y_train))
print('TrainModel Start!')
pickle.dump(clf, open(MODELPATH + '/NB_zhuyuan.pickle', 'wb'), -1)
print('TrainModel Finish!')
return clf
def trainModel(trainFile,testFile):
import numpy as np
import TT_ML.data_helper.data_prepare as dataProcess
data_helper=dataProcess.DataPrepare()
x_train,y_train,x_test,y_test=data_helper.loadLibSVMFile(trainFile,testFile)
from sklearn.linear_model import SGDClassifier
clf=SGDClassifier(loss='log',penalty='l2')#modified_huber
clf.partial_fit(x_train,y_train,classes=np.unique(y_train))
#from sklearn.neighbors import KNeighborsClassifier
#clf=KNeighborsClassifier(n_neighbors=10)
clf.fit(x_train,y_train)
print('TrainModel Start!')
pickle.dump(clf, open('./SGD_zhuyuan.pickle', 'wb'), -1)
print('TrainModel Finish!')
return clf
def loadModel_Kears():
from keras.models import load_model
model = load_model('my_model.h5')
import TT_ML.data_helper.data_prepare as dataProcess
data_helper = dataProcess.DataPrepare()
x_train, y_train, x_test, y_test = data_helper.loadLibSVMFile(
'/home/jq/jeeker/zy_data2/train.libsvm',
'/home/jq/jeeker/zy_data2/test.libsvm')
y_pred = model.predict(x_train[:60000], batch_size=256)
import numpy as np
note_prediction = np.argmax(y_pred, axis=1)
print('*' * 20, note_prediction, '*' * 20)
from sklearn.metrics import classification_report, confusion_matrix
#print(confusion_matrix(y_test, note_prediction))
print(classification_report(y_train[:60000], note_prediction))
def trainModel(trainFile, testFile):
import numpy as np
import TT_ML.data_helper.data_prepare as dataProcess
data_helper = dataProcess.DataPrepare()
#加载训练集和测试集
x_train, y_train, x_test, y_test = data_helper.loadLibSVMFile(
trainFile, testFile)
#设置模型的参数，调参
import TT_ML.singleLearning.KNN as KNN
clf = KNN.KNN_normal(x_train, y_train, x_test, y_test)
#训练模型
print('TrainModel Start!')
#保存模型,线上部署需要
pickle.dump(clf, open(MODELPATH + '/KNN_zhuyuan.pickle', 'wb'), -1)
print('TrainModel Finish!')
return clf
def trainModel(trainFile,testFile):
import numpy as np
import TT_ML.data_helper.data_prepare as dataProcess
import TT_ML.singleLearning.NB as NB
data_helper=dataProcess.DataPrepare()
#加载训练集和测试集
x_train,y_train,x_test,y_test=data_helper.loadLibSVMFile(trainFile,testFile)
#设置模型的参数，调参
clf = NB.NB_normal(x_train,y_train,x_test,y_test,modelName='BernoulliNB')
#训练模型
#clf.fit(x_train,y_train)
print('TrainModel Start!')
#保存模型,线上部署需要
pickle.dump(clf, open(MODELPATH+'/NB_zhuyuan.pickle', 'wb'), -1)
print('TrainModel Finish!')
return clf
def trainModel(trainFile, testFile):
import numpy as np
import TT_ML.data_helper.data_prepare as dataProcess
data_helper = dataProcess.DataPrepare()
#加载训练集和测试集
x_train, y_train, x_test, y_test = data_helper.loadLibSVMFile(
trainFile, testFile)
from sklearn.linear_model import SGDClassifier
#设置模型的参数，调参
import TT_ML.ensembleLearning.RandomForest as ensemble_RF
clf = ensemble_RF.RF_normal(x_train, y_train, x_test, y_test)
#训练模型
print('TrainModel Start!')
#保存模型,线上部署需要
pickle.dump(clf, open(MODELPATH + '/RF_zhuyuan.pickle', 'wb'), -1)
print('TrainModel Finish!')
return clf
def trainModel(trainFile,testFile):
import numpy as np
import TT_ML.data_helper.data_prepare as dataProcess
data_helper=dataProcess.DataPrepare()
x_train,y_train,x_test,y_test=data_helper.loadLibSVMFile(trainFile,testFile)
from sklearn.naive_bayes import MultinomialNB
clf = MultinomialNB()
print('交叉验证开始')
from sklearn.model_selection import ShuffleSplit
from sklearn.model_selection import cross_val_score
cv = ShuffleSplit(n_splits=5, test_size=0.2, random_state=0)
accs=cross_val_score(clf,x_train,y_train,cv=cv,n_jobs=5)
print('交叉验证结果:',accs)
print('交叉验证结束')
return clf
def trainModel(trainFile, testFile):
import numpy as np
import TT_ML.data_helper.data_prepare as dataProcess
data_helper = dataProcess.DataPrepare()
x_train, y_train, x_test, y_test = data_helper.loadLibSVMFile(
trainFile, testFile)
from sklearn.ensemble import RandomForestClassifier
clf = RandomForestClassifier(n_estimators=70, class_weight='balanced')
print('TrainModel Finish!')
print('交叉验证开始')
from sklearn.model_selection import ShuffleSplit
from sklearn.model_selection import cross_val_score
cv = ShuffleSplit(n_splits=5, test_size=0.01, random_state=0)
accs = cross_val_score(clf, x_train, y_train, cv=cv, n_jobs=5)
print('交叉验证结果:', accs)
print('交叉验证结束')
return clf
def trainModel(trainFile, testFile):
import numpy as np
import TT_ML.data_helper.data_prepare as dataProcess
data_helper = dataProcess.DataPrepare()
#加载训练集和测试集
x_train, y_train, x_test, y_test = data_helper.loadLibSVMFile(
trainFile, testFile)
from sklearn.linear_model import SGDClassifier
#设置模型的参数，调参
clf = SGDClassifier(loss='log', penalty='l2',
class_weight='balanced') #modified_huber
#clf.partial_fit(x_train,y_train,classes=np.unique(y_train))
#训练模型
clf.fit(x_train, y_train)
print('TrainModel Start!')
#保存模型,线上部署需要
pickle.dump(clf, open(MODELPATH + '/SGD_zhuyuan.pickle', 'wb'), -1)
print('TrainModel Finish!')
return clf
def trainModel(trainFile, testFile):
import numpy as np
import TT_ML.data_helper.data_prepare as dataProcess
data_helper = dataProcess.DataPrepare()
x_train, y_train, x_test, y_test = data_helper.loadLibSVMFile(
trainFile, testFile)
#from sklearn.linear_model import SGDClassifier
#clf=SGDClassifier(loss='log',penalty='l2',class_weight='balanced')#modified_huber
#clf.partial_fit(x_train,y_train,classes=np.unique(y_train))
#clf.fit(x_train,y_train)
#from sklearn.neighbors import KNeighborsClassifier
#clf=KNeighborsClassifier(n_neighbors=10)
#clf.fit(x_train,y_train)
from sklearn.ensemble import RandomForestClassifier
clf = RandomForestClassifier(n_estimators=70, class_weight='balanced')
clf = clf.fit(x_train, y_train)
print('TrainModel Start!')
pickle.dump(clf, open(MODELPATH + '/RF_zhuyuan.pickle', 'wb'), -1)
print('TrainModel Finish!')
return clf