Kaggle Brush the game's sharp weapon, lr,lgbm,xgboost,keras__ machine learning

Brush the Race tool, thank the people who share.

Summary

Recently played a variety of games, here to share some general Model, a little change can be used

Environment: Python 3.5.2

Xgboost: http://blog.csdn.net/han_xiaoyang/article/details/52665396
Xgboost Official API:
Http://xgboost.readthedocs.io/en/latest//python/python_api.htmlpreprocess[Python] View plain copy # Common preprocessing framework import pandas as PD import NumPy as NP import scipy as SP # file Read Def           Read_csv_file (F, Logging=false): Print ("========== read Data =========") data = Pd.read_csv (f) If logging: Print (Data.head (5)) print (F, "contains the following") print (data.columns.values) print (data.describe ()) Print (Data.info ()) Return data

Logistic regression [python]  View Plain  copy #  Universal logisticregression Framework       Import pandas  as pd   import numpy as np   from scipy import  sparse   from sklearn.preprocessing import onehotencoder   from  sklearn.linear_model import logisticregression   from sklearn.preprocessing  import standardscaler      # 1. load data   df_train =  PD. Dataframe ()    df_test  = pd. Dataframe ()    y_train = df_train[' label '].values      # 2.  process data   Ss = standardscaler ()          # 3.  feature engineering/encoding   # 3.1 for labeled feature   Enc = onehotencoder ()    feats = ["CreativEID ", " Adid ", " Campaignid "]   for i, feat in enumerate (feats):        x_train = enc.fit_transform (Df_train[feat].values.reshape ( -1, 1))        x_test = enc.fit_transform (Df_test[feat].values.reshape (-1),  1))