wonderful
/
kaggle
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

修改rf-python3.6.py

This commit is contained in:
wang-sw 2018-05-19 19:41:26 +08:00
parent ac59701ca7
commit f1e4c6073b
1 changed files with 9 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
src/python/getting-started/digit-recognizer/rf-python3.6.py
View File

@ -54,7 +54,7 @@ def dRPCA(data, COMPONENT_NUM=100):
# 训练模型 # 训练模型
def trainModel(Xtrain, xtest): def trainModel(X_train, y_train):
print('Train RF...') print('Train RF...')
clf = RandomForestClassifier( clf = RandomForestClassifier(
n_estimators=140, n_estimators=140,
@ -62,13 +62,13 @@ def trainModel(Xtrain, xtest):
min_samples_split=2, min_samples_split=2,
min_samples_leaf=1, min_samples_leaf=1,
random_state=34) random_state=34)
clf.fit(Xtrain, xtest) # 训练rf clf.fit(X_train, y_train) # 训练rf
# clf=LGBMClassifier(num_leaves=63, max_depth=7, n_estimators=80, n_jobs=20) # clf=LGBMClassifier(num_leaves=63, max_depth=7, n_estimators=80, n_jobs=20)
# param_test1 = {'n_estimators':arange(10,150,10),'max_depth':arange(1,21,1)} # param_test1 = {'n_estimators':arange(10,150,10),'max_depth':arange(1,21,1)}
# gsearch1 = GridSearchCV(estimator = clf, param_grid = param_test1, scoring='accuracy',iid=False,cv=5) # gsearch1 = GridSearchCV(estimator = clf, param_grid = param_test1, scoring='accuracy',iid=False,cv=5)
# gsearch1.fit(Xtrain,xtest) # gsearch1.fit(X_train, y_train)
# print(gsearch1.grid_scores_, gsearch1.best_params_, gsearch1.best_score_) # print(gsearch1.grid_scores_, gsearch1.best_params_, gsearch1.best_score_)
# clf=gsearch1.best_estimator_ # clf=gsearch1.best_estimator_
@ -76,8 +76,8 @@ def trainModel(Xtrain, xtest):
# 计算准确率 # 计算准确率
def printAccuracy(ytest ,y_predict): def printAccuracy(y_test ,y_predict):
zeroLable = ytest - y_predict zeroLable = y_test - y_predict
rightCount = 0 rightCount = 0
for i in range(len(zeroLable)): for i in range(len(zeroLable)):
if list(zeroLable)[i] == 0: if list(zeroLable)[i] == 0:
@ -123,18 +123,18 @@ def trainRF():
# 模型训练 (数据预处理-降维) # 模型训练 (数据预处理-降维)
data_pca = dRPCA(data,100) data_pca = dRPCA(data,100)
Xtrain, Ytrain, xtest, ytest = train_test_split( X_train, X_test, y_train, y_test = train_test_split(
data_pca[0:len(train_data)], label, test_size=0.1, random_state=34) data_pca[0:len(train_data)], label, test_size=0.1, random_state=34)
rfClf = trainModel(Xtrain, xtest) rfClf = trainModel(X_train, y_train)
# 保存结果 # 保存结果
storeModel(data_pca[len(train_data):], os.path.join(data_dir, 'output/Result_sklearn_rf.pcaPreData')) storeModel(data_pca[len(train_data):], os.path.join(data_dir, 'output/Result_sklearn_rf.pcaPreData'))
storeModel(rfClf, os.path.join(data_dir, 'output/Result_sklearn_rf.model')) storeModel(rfClf, os.path.join(data_dir, 'output/Result_sklearn_rf.model'))
# 模型准确率 # 模型准确率
y_predict = rfClf.predict(Ytrain) y_predict = rfClf.predict(X_test)
printAccuracy(ytest, y_predict) printAccuracy(y_test, y_predict)
print("finish!") print("finish!")
stopTime = time.time() stopTime = time.time()