特征工程优化

train/train_service.py

@@ -138,7 +138,7 @@ def train(emb_columns, number_columns, oneHot_columns, inputs, train_dataset):
 
     # train the model
     print("train start...")
-    model.fit(train_dataset, epochs=10)
+    model.fit(train_dataset, epochs=5)
     print("train end...")
     print("train save...")
 
@@ -177,10 +177,6 @@ def addDays(n, format="%Y%m%d"):
 
 if __name__ == '__main__':
 
-    isTest = False
-    if (len(sys.argv) >= 2 and "test" == sys.argv[1]):
-        isTest = True
-
     curTime = time.strftime("%Y-%m-%d %H:%M:%S", time.localtime())
     print("train_service执行时间:{}".format(curTime))
     splitTimestamp = int(time.mktime(time.strptime(addDays(-1), "%Y%m%d")))
@@ -195,8 +191,7 @@ if __name__ == '__main__':
     print(df_train.dtypes)
 
     df_test = df_train.loc[df_train['timestamp']>=splitTimestamp]
-    if isTest:
-        df_train = df_train.loc[df_train['timestamp'] < splitTimestamp]
+    df_train = df_train.loc[df_train['timestamp'] < splitTimestamp]
     # df_test = loadData(data_path_test)
     timestmp2 = int(round(time.time()))
     print("读取数据耗时s:{}".format(timestmp2 - timestmp1))