In [1]:

    

# imports
import h2o
from h2o.estimators.xgboost import H2OXGBoostEstimator
import matplotlib.pyplot as plt
%matplotlib inline
import numpy as np
import pandas as pd
import xgboost as xgb


    

In [2]:

    

# start h2o
h2o.init()
h2o.remove_all()


    

    




Checking whether there is an H2O instance running at http://localhost:54321..... not found.
Attempting to start a local H2O server...
; Java HotSpot(TM) 64-Bit Server VM (build 24.79-b02, mixed mode)
  Starting server from C:\Anaconda3\lib\site-packages\h2o\backend\bin\h2o.jar
  Ice root: C:\Users\p\AppData\Local\Temp\tmp92vc54ht
  JVM stdout: C:\Users\p\AppData\Local\Temp\tmp92vc54ht\h2o_p_started_from_python.out
  JVM stderr: C:\Users\p\AppData\Local\Temp\tmp92vc54ht\h2o_p_started_from_python.err
  Server is running at http://127.0.0.1:54321
Connecting to H2O server at http://127.0.0.1:54321... successful.






    





H2O cluster uptime:
01 secs
H2O cluster version:
3.10.5.4
H2O cluster version age:
6 days 
H2O cluster name:
H2O_from_python_p_q2okdx
H2O cluster total nodes:
1
H2O cluster free memory:
3.538 Gb
H2O cluster total cores:
8
H2O cluster allowed cores:
8
H2O cluster status:
accepting new members, healthy
H2O connection url:
http://127.0.0.1:54321
H2O connection proxy:
None
H2O internal security:
False
Python version:
3.5.2 final

In [3]:

    

# load clean data
path = '../../03_regression/data/train.csv'
frame = h2o.import_file(path=path)


    

    




Parse progress: |█████████████████████████████████████████████████████████| 100%

In [4]:

    

# assign target and inputs
y = 'SalePrice'
X = [name for name in frame.columns if name not in [y, 'Id']]


    

In [5]:

    

# determine column types
# impute
reals, enums = [], []
for key, val in frame.types.items():
    if key in X:
        if val == 'enum':
            enums.append(key)
        else: 
            reals.append(key)
            
_ = frame[reals].impute(method='median')


    

In [6]:

    

# split into training and validation
train, valid = frame.split_frame([0.7], seed=12345)


    

In [7]:

    

# for convenience create a tuple for xgboost monotone_constraints parameter
mono_constraints = tuple(int(i) for i in np.ones(shape=(int(1), len(reals))).tolist()[0])


    

In [8]:

    

# Check log transform - looks good
%matplotlib inline
train['SalePrice'].log().as_data_frame().hist()

# Execute log transform
train['SalePrice'] = train['SalePrice'].log()
valid['SalePrice'] = valid['SalePrice'].log()
print(train[0:3, 'SalePrice'])


    

    







  SalePrice


    12.2477
    12.109 
    12.3172








    










    

In [9]:

    

ave_y = train['SalePrice'].mean()[0]

# XGBoost uses SVMLight data structure, not Numpy arrays or Pandas data frames 
dtrain = xgb.DMatrix(train.as_data_frame()[reals],
                      train.as_data_frame()['SalePrice'])
dvalid = xgb.DMatrix(valid.as_data_frame()[reals],
                      valid.as_data_frame()['SalePrice'])

# tuning parameters
params = {
    'objective': 'reg:linear',
    'booster': 'gbtree', 
    'eval_metric': 'rmse',
    'eta': 0.005,
    'subsample': 0.1, 
    'colsample_bytree': 0.8,
    'max_depth': 5,
    'reg_alpha' : 0.01,
    'reg_lambda' : 0.0,
    'monotone_constraints':mono_constraints,
    'base_score': ave_y,
    'silent': 0,
    'seed': 12345,
}

# watchlist is used for early stopping
watchlist = [(dtrain, 'train'), (dvalid, 'eval')]

# train model
xgb_model1 = xgb.train(params, 
                       dtrain, 
                       1000,
                       evals=watchlist, 
                       early_stopping_rounds=50, 
                       verbose_eval=True)


    

    




[0]	train-rmse:0.3924	eval-rmse:0.410023
Multiple eval metrics have been passed: 'eval-rmse' will be used for early stopping.

Will train until eval-rmse hasn't improved in 50 rounds.
[1]	train-rmse:0.390954	eval-rmse:0.408467
[2]	train-rmse:0.389618	eval-rmse:0.407106
[3]	train-rmse:0.388497	eval-rmse:0.405756
[4]	train-rmse:0.387136	eval-rmse:0.404236
[5]	train-rmse:0.385714	eval-rmse:0.402831
[6]	train-rmse:0.384449	eval-rmse:0.401643
[7]	train-rmse:0.382743	eval-rmse:0.399816
[8]	train-rmse:0.381471	eval-rmse:0.398484
[9]	train-rmse:0.380089	eval-rmse:0.397062
[10]	train-rmse:0.378992	eval-rmse:0.395882
[11]	train-rmse:0.377797	eval-rmse:0.394543
[12]	train-rmse:0.376604	eval-rmse:0.39328
[13]	train-rmse:0.37533	eval-rmse:0.391908
[14]	train-rmse:0.373751	eval-rmse:0.390438
[15]	train-rmse:0.37246	eval-rmse:0.389129
[16]	train-rmse:0.371113	eval-rmse:0.38778
[17]	train-rmse:0.369844	eval-rmse:0.386449
[18]	train-rmse:0.368511	eval-rmse:0.385084
[19]	train-rmse:0.36718	eval-rmse:0.383684
[20]	train-rmse:0.365808	eval-rmse:0.382153
[21]	train-rmse:0.364398	eval-rmse:0.380706
[22]	train-rmse:0.363094	eval-rmse:0.379213
[23]	train-rmse:0.361733	eval-rmse:0.377871
[24]	train-rmse:0.360402	eval-rmse:0.376537
[25]	train-rmse:0.359047	eval-rmse:0.37518
[26]	train-rmse:0.357816	eval-rmse:0.373909
[27]	train-rmse:0.356583	eval-rmse:0.372763
[28]	train-rmse:0.355392	eval-rmse:0.371539
[29]	train-rmse:0.354017	eval-rmse:0.370001
[30]	train-rmse:0.352962	eval-rmse:0.368869
[31]	train-rmse:0.351735	eval-rmse:0.367572
[32]	train-rmse:0.350414	eval-rmse:0.366278
[33]	train-rmse:0.34909	eval-rmse:0.364852
[34]	train-rmse:0.348044	eval-rmse:0.363857
[35]	train-rmse:0.346958	eval-rmse:0.362709
[36]	train-rmse:0.345771	eval-rmse:0.361432
[37]	train-rmse:0.344657	eval-rmse:0.360253
[38]	train-rmse:0.343384	eval-rmse:0.358927
[39]	train-rmse:0.342106	eval-rmse:0.357613
[40]	train-rmse:0.340843	eval-rmse:0.356215
[41]	train-rmse:0.339678	eval-rmse:0.354896
[42]	train-rmse:0.338493	eval-rmse:0.353594
[43]	train-rmse:0.337334	eval-rmse:0.352363
[44]	train-rmse:0.336152	eval-rmse:0.35105
[45]	train-rmse:0.334906	eval-rmse:0.34972
[46]	train-rmse:0.333908	eval-rmse:0.348785
[47]	train-rmse:0.332783	eval-rmse:0.347465
[48]	train-rmse:0.331776	eval-rmse:0.346471
[49]	train-rmse:0.330468	eval-rmse:0.345205
[50]	train-rmse:0.329272	eval-rmse:0.343932
[51]	train-rmse:0.328072	eval-rmse:0.342741
[52]	train-rmse:0.326961	eval-rmse:0.341641
[53]	train-rmse:0.325797	eval-rmse:0.340275
[54]	train-rmse:0.324647	eval-rmse:0.339084
[55]	train-rmse:0.323357	eval-rmse:0.337551
[56]	train-rmse:0.32236	eval-rmse:0.336513
[57]	train-rmse:0.321196	eval-rmse:0.335334
[58]	train-rmse:0.320245	eval-rmse:0.33428
[59]	train-rmse:0.319255	eval-rmse:0.333335
[60]	train-rmse:0.318107	eval-rmse:0.331977
[61]	train-rmse:0.316973	eval-rmse:0.330865
[62]	train-rmse:0.315922	eval-rmse:0.329654
[63]	train-rmse:0.314727	eval-rmse:0.328422
[64]	train-rmse:0.313633	eval-rmse:0.327252
[65]	train-rmse:0.312645	eval-rmse:0.326283
[66]	train-rmse:0.311411	eval-rmse:0.324992
[67]	train-rmse:0.310406	eval-rmse:0.323971
[68]	train-rmse:0.309295	eval-rmse:0.322864
[69]	train-rmse:0.3084	eval-rmse:0.321954
[70]	train-rmse:0.307538	eval-rmse:0.321043
[71]	train-rmse:0.306586	eval-rmse:0.319932
[72]	train-rmse:0.305433	eval-rmse:0.318684
[73]	train-rmse:0.304577	eval-rmse:0.317955
[74]	train-rmse:0.303391	eval-rmse:0.316781
[75]	train-rmse:0.302433	eval-rmse:0.315815
[76]	train-rmse:0.301417	eval-rmse:0.314746
[77]	train-rmse:0.300425	eval-rmse:0.313772
[78]	train-rmse:0.299632	eval-rmse:0.312921
[79]	train-rmse:0.298593	eval-rmse:0.311855
[80]	train-rmse:0.297476	eval-rmse:0.310512
[81]	train-rmse:0.296454	eval-rmse:0.309488
[82]	train-rmse:0.295476	eval-rmse:0.308448
[83]	train-rmse:0.294545	eval-rmse:0.307467
[84]	train-rmse:0.293698	eval-rmse:0.306618
[85]	train-rmse:0.292919	eval-rmse:0.3058
[86]	train-rmse:0.291839	eval-rmse:0.304714
[87]	train-rmse:0.29112	eval-rmse:0.30394
[88]	train-rmse:0.290434	eval-rmse:0.303167
[89]	train-rmse:0.289413	eval-rmse:0.301853
[90]	train-rmse:0.288591	eval-rmse:0.300972
[91]	train-rmse:0.287797	eval-rmse:0.3003
[92]	train-rmse:0.287085	eval-rmse:0.299615
[93]	train-rmse:0.286334	eval-rmse:0.298834
[94]	train-rmse:0.285387	eval-rmse:0.297891
[95]	train-rmse:0.284492	eval-rmse:0.296876
[96]	train-rmse:0.283575	eval-rmse:0.29586
[97]	train-rmse:0.282698	eval-rmse:0.295007
[98]	train-rmse:0.281767	eval-rmse:0.293973
[99]	train-rmse:0.280898	eval-rmse:0.293025
[100]	train-rmse:0.279944	eval-rmse:0.291954
[101]	train-rmse:0.278922	eval-rmse:0.290909
[102]	train-rmse:0.277925	eval-rmse:0.289707
[103]	train-rmse:0.277078	eval-rmse:0.288749
[104]	train-rmse:0.276306	eval-rmse:0.287823
[105]	train-rmse:0.275496	eval-rmse:0.287069
[106]	train-rmse:0.274698	eval-rmse:0.286221
[107]	train-rmse:0.273866	eval-rmse:0.285274
[108]	train-rmse:0.273091	eval-rmse:0.284503
[109]	train-rmse:0.272268	eval-rmse:0.283633
[110]	train-rmse:0.27163	eval-rmse:0.282928
[111]	train-rmse:0.270892	eval-rmse:0.282141
[112]	train-rmse:0.270112	eval-rmse:0.281381
[113]	train-rmse:0.269582	eval-rmse:0.280827
[114]	train-rmse:0.268758	eval-rmse:0.280047
[115]	train-rmse:0.267932	eval-rmse:0.279143
[116]	train-rmse:0.267085	eval-rmse:0.278244
[117]	train-rmse:0.266338	eval-rmse:0.277479
[118]	train-rmse:0.265643	eval-rmse:0.276677
[119]	train-rmse:0.264955	eval-rmse:0.275966
[120]	train-rmse:0.26422	eval-rmse:0.275228
[121]	train-rmse:0.263567	eval-rmse:0.274551
[122]	train-rmse:0.26288	eval-rmse:0.273894
[123]	train-rmse:0.262087	eval-rmse:0.273105
[124]	train-rmse:0.261296	eval-rmse:0.272233
[125]	train-rmse:0.260619	eval-rmse:0.271463
[126]	train-rmse:0.259946	eval-rmse:0.270694
[127]	train-rmse:0.259205	eval-rmse:0.269923
[128]	train-rmse:0.258388	eval-rmse:0.269018
[129]	train-rmse:0.257643	eval-rmse:0.268309
[130]	train-rmse:0.256697	eval-rmse:0.267332
[131]	train-rmse:0.255963	eval-rmse:0.26669
[132]	train-rmse:0.255326	eval-rmse:0.265973
[133]	train-rmse:0.254567	eval-rmse:0.265155
[134]	train-rmse:0.253865	eval-rmse:0.264362
[135]	train-rmse:0.25307	eval-rmse:0.263556
[136]	train-rmse:0.252377	eval-rmse:0.262872
[137]	train-rmse:0.251536	eval-rmse:0.26203
[138]	train-rmse:0.250944	eval-rmse:0.261465
[139]	train-rmse:0.250208	eval-rmse:0.260754
[140]	train-rmse:0.249587	eval-rmse:0.260099
[141]	train-rmse:0.248974	eval-rmse:0.259413
[142]	train-rmse:0.248352	eval-rmse:0.258791
[143]	train-rmse:0.247759	eval-rmse:0.258198
[144]	train-rmse:0.247017	eval-rmse:0.257496
[145]	train-rmse:0.246287	eval-rmse:0.256718
[146]	train-rmse:0.245688	eval-rmse:0.256152
[147]	train-rmse:0.244938	eval-rmse:0.255274
[148]	train-rmse:0.244268	eval-rmse:0.254508
[149]	train-rmse:0.243695	eval-rmse:0.253796
[150]	train-rmse:0.243007	eval-rmse:0.252967
[151]	train-rmse:0.242488	eval-rmse:0.252324
[152]	train-rmse:0.241913	eval-rmse:0.251672
[153]	train-rmse:0.241123	eval-rmse:0.250736
[154]	train-rmse:0.240559	eval-rmse:0.250131
[155]	train-rmse:0.239988	eval-rmse:0.249553
[156]	train-rmse:0.239331	eval-rmse:0.248683
[157]	train-rmse:0.23872	eval-rmse:0.247863
[158]	train-rmse:0.237997	eval-rmse:0.247032
[159]	train-rmse:0.237253	eval-rmse:0.246289
[160]	train-rmse:0.236633	eval-rmse:0.245547
[161]	train-rmse:0.236091	eval-rmse:0.244908
[162]	train-rmse:0.235566	eval-rmse:0.244327
[163]	train-rmse:0.23486	eval-rmse:0.243538
[164]	train-rmse:0.234117	eval-rmse:0.242741
[165]	train-rmse:0.233493	eval-rmse:0.242019
[166]	train-rmse:0.232785	eval-rmse:0.241236
[167]	train-rmse:0.232113	eval-rmse:0.240538
[168]	train-rmse:0.231448	eval-rmse:0.239996
[169]	train-rmse:0.230898	eval-rmse:0.239428
[170]	train-rmse:0.230326	eval-rmse:0.238819
[171]	train-rmse:0.229715	eval-rmse:0.238149
[172]	train-rmse:0.229058	eval-rmse:0.237391
[173]	train-rmse:0.228336	eval-rmse:0.236686
[174]	train-rmse:0.227743	eval-rmse:0.236005
[175]	train-rmse:0.227188	eval-rmse:0.235316
[176]	train-rmse:0.226576	eval-rmse:0.234718
[177]	train-rmse:0.225921	eval-rmse:0.233964
[178]	train-rmse:0.225286	eval-rmse:0.233209
[179]	train-rmse:0.224611	eval-rmse:0.232514
[180]	train-rmse:0.224121	eval-rmse:0.231974
[181]	train-rmse:0.223554	eval-rmse:0.231389
[182]	train-rmse:0.222938	eval-rmse:0.230591
[183]	train-rmse:0.222368	eval-rmse:0.230006
[184]	train-rmse:0.221852	eval-rmse:0.229488
[185]	train-rmse:0.221213	eval-rmse:0.22879
[186]	train-rmse:0.220518	eval-rmse:0.227971
[187]	train-rmse:0.220076	eval-rmse:0.227468
[188]	train-rmse:0.219669	eval-rmse:0.227056
[189]	train-rmse:0.21918	eval-rmse:0.226464
[190]	train-rmse:0.21871	eval-rmse:0.225987
[191]	train-rmse:0.218218	eval-rmse:0.225402
[192]	train-rmse:0.217701	eval-rmse:0.224801
[193]	train-rmse:0.217214	eval-rmse:0.224224
[194]	train-rmse:0.216694	eval-rmse:0.223686
[195]	train-rmse:0.216036	eval-rmse:0.223017
[196]	train-rmse:0.215364	eval-rmse:0.222351
[197]	train-rmse:0.214899	eval-rmse:0.221838
[198]	train-rmse:0.214338	eval-rmse:0.221244
[199]	train-rmse:0.213877	eval-rmse:0.220764
[200]	train-rmse:0.213324	eval-rmse:0.220193
[201]	train-rmse:0.212605	eval-rmse:0.219451
[202]	train-rmse:0.212054	eval-rmse:0.218773
[203]	train-rmse:0.211368	eval-rmse:0.218094
[204]	train-rmse:0.210884	eval-rmse:0.217538
[205]	train-rmse:0.210421	eval-rmse:0.217114
[206]	train-rmse:0.209907	eval-rmse:0.216516
[207]	train-rmse:0.209387	eval-rmse:0.215891
[208]	train-rmse:0.208686	eval-rmse:0.214936
[209]	train-rmse:0.208159	eval-rmse:0.214348
[210]	train-rmse:0.207863	eval-rmse:0.214075
[211]	train-rmse:0.207414	eval-rmse:0.213616
[212]	train-rmse:0.206923	eval-rmse:0.213013
[213]	train-rmse:0.206341	eval-rmse:0.212406
[214]	train-rmse:0.205881	eval-rmse:0.211968
[215]	train-rmse:0.205259	eval-rmse:0.21123
[216]	train-rmse:0.204775	eval-rmse:0.210702
[217]	train-rmse:0.20435	eval-rmse:0.21011
[218]	train-rmse:0.203849	eval-rmse:0.209552
[219]	train-rmse:0.203468	eval-rmse:0.209204
[220]	train-rmse:0.202964	eval-rmse:0.208647
[221]	train-rmse:0.20258	eval-rmse:0.208229
[222]	train-rmse:0.202187	eval-rmse:0.207791
[223]	train-rmse:0.201756	eval-rmse:0.20726
[224]	train-rmse:0.201325	eval-rmse:0.206733
[225]	train-rmse:0.200918	eval-rmse:0.206284
[226]	train-rmse:0.200547	eval-rmse:0.205834
[227]	train-rmse:0.200003	eval-rmse:0.205323
[228]	train-rmse:0.199743	eval-rmse:0.204965
[229]	train-rmse:0.199318	eval-rmse:0.204527
[230]	train-rmse:0.19873	eval-rmse:0.203787
[231]	train-rmse:0.198261	eval-rmse:0.203157
[232]	train-rmse:0.197734	eval-rmse:0.20245
[233]	train-rmse:0.197388	eval-rmse:0.202029
[234]	train-rmse:0.197017	eval-rmse:0.20157
[235]	train-rmse:0.196598	eval-rmse:0.201135
[236]	train-rmse:0.196114	eval-rmse:0.200748
[237]	train-rmse:0.195742	eval-rmse:0.200389
[238]	train-rmse:0.195308	eval-rmse:0.199833
[239]	train-rmse:0.194935	eval-rmse:0.199443
[240]	train-rmse:0.194592	eval-rmse:0.198978
[241]	train-rmse:0.193996	eval-rmse:0.198277
[242]	train-rmse:0.193602	eval-rmse:0.197903
[243]	train-rmse:0.193149	eval-rmse:0.197439
[244]	train-rmse:0.192726	eval-rmse:0.196872
[245]	train-rmse:0.192395	eval-rmse:0.196441
[246]	train-rmse:0.19214	eval-rmse:0.196092
[247]	train-rmse:0.191754	eval-rmse:0.19566
[248]	train-rmse:0.191368	eval-rmse:0.195321
[249]	train-rmse:0.190934	eval-rmse:0.194774
[250]	train-rmse:0.190518	eval-rmse:0.194258
[251]	train-rmse:0.190115	eval-rmse:0.193791
[252]	train-rmse:0.189729	eval-rmse:0.193406
[253]	train-rmse:0.189412	eval-rmse:0.193034
[254]	train-rmse:0.188999	eval-rmse:0.192507
[255]	train-rmse:0.188653	eval-rmse:0.192147
[256]	train-rmse:0.188278	eval-rmse:0.191794
[257]	train-rmse:0.187767	eval-rmse:0.191309
[258]	train-rmse:0.187474	eval-rmse:0.191003
[259]	train-rmse:0.18708	eval-rmse:0.19056
[260]	train-rmse:0.186751	eval-rmse:0.190134
[261]	train-rmse:0.18632	eval-rmse:0.18964
[262]	train-rmse:0.185957	eval-rmse:0.189304
[263]	train-rmse:0.185497	eval-rmse:0.188872
[264]	train-rmse:0.18518	eval-rmse:0.188537
[265]	train-rmse:0.184851	eval-rmse:0.188166
[266]	train-rmse:0.184492	eval-rmse:0.187696
[267]	train-rmse:0.184162	eval-rmse:0.187351
[268]	train-rmse:0.183756	eval-rmse:0.186909
[269]	train-rmse:0.183425	eval-rmse:0.186566
[270]	train-rmse:0.183157	eval-rmse:0.186233
[271]	train-rmse:0.182845	eval-rmse:0.185848
[272]	train-rmse:0.182382	eval-rmse:0.185321
[273]	train-rmse:0.181988	eval-rmse:0.184951
[274]	train-rmse:0.18166	eval-rmse:0.184534
[275]	train-rmse:0.181327	eval-rmse:0.184128
[276]	train-rmse:0.180906	eval-rmse:0.183577
[277]	train-rmse:0.180485	eval-rmse:0.183009
[278]	train-rmse:0.18008	eval-rmse:0.182563
[279]	train-rmse:0.179739	eval-rmse:0.182097
[280]	train-rmse:0.179393	eval-rmse:0.181688
[281]	train-rmse:0.179023	eval-rmse:0.181365
[282]	train-rmse:0.178572	eval-rmse:0.180867
[283]	train-rmse:0.178326	eval-rmse:0.18053
[284]	train-rmse:0.178034	eval-rmse:0.180175
[285]	train-rmse:0.177641	eval-rmse:0.17966
[286]	train-rmse:0.177327	eval-rmse:0.179176
[287]	train-rmse:0.176989	eval-rmse:0.17889
[288]	train-rmse:0.176747	eval-rmse:0.178603
[289]	train-rmse:0.176487	eval-rmse:0.178302
[290]	train-rmse:0.17609	eval-rmse:0.177954
[291]	train-rmse:0.175773	eval-rmse:0.177582
[292]	train-rmse:0.175445	eval-rmse:0.177289
[293]	train-rmse:0.175207	eval-rmse:0.177089
[294]	train-rmse:0.174892	eval-rmse:0.176631
[295]	train-rmse:0.174567	eval-rmse:0.176259
[296]	train-rmse:0.174219	eval-rmse:0.175851
[297]	train-rmse:0.173919	eval-rmse:0.175445
[298]	train-rmse:0.173607	eval-rmse:0.175114
[299]	train-rmse:0.173319	eval-rmse:0.174804
[300]	train-rmse:0.17305	eval-rmse:0.174532
[301]	train-rmse:0.172741	eval-rmse:0.174241
[302]	train-rmse:0.172493	eval-rmse:0.173963
[303]	train-rmse:0.172104	eval-rmse:0.173566
[304]	train-rmse:0.171852	eval-rmse:0.173254
[305]	train-rmse:0.171537	eval-rmse:0.172841
[306]	train-rmse:0.171339	eval-rmse:0.172652
[307]	train-rmse:0.170964	eval-rmse:0.172193
[308]	train-rmse:0.170696	eval-rmse:0.171894
[309]	train-rmse:0.170394	eval-rmse:0.171555
[310]	train-rmse:0.170058	eval-rmse:0.17119
[311]	train-rmse:0.169682	eval-rmse:0.170843
[312]	train-rmse:0.1693	eval-rmse:0.170432
[313]	train-rmse:0.169042	eval-rmse:0.170165
[314]	train-rmse:0.168781	eval-rmse:0.169887
[315]	train-rmse:0.168508	eval-rmse:0.169491
[316]	train-rmse:0.168253	eval-rmse:0.169193
[317]	train-rmse:0.167968	eval-rmse:0.16884
[318]	train-rmse:0.167764	eval-rmse:0.168594
[319]	train-rmse:0.167357	eval-rmse:0.1682
[320]	train-rmse:0.167047	eval-rmse:0.167865
[321]	train-rmse:0.1668	eval-rmse:0.16759
[322]	train-rmse:0.166481	eval-rmse:0.167166
[323]	train-rmse:0.166182	eval-rmse:0.166894
[324]	train-rmse:0.166005	eval-rmse:0.166694
[325]	train-rmse:0.165763	eval-rmse:0.166461
[326]	train-rmse:0.165443	eval-rmse:0.166138
[327]	train-rmse:0.165182	eval-rmse:0.165808
[328]	train-rmse:0.16489	eval-rmse:0.165487
[329]	train-rmse:0.16466	eval-rmse:0.165233
[330]	train-rmse:0.16438	eval-rmse:0.164916
[331]	train-rmse:0.164039	eval-rmse:0.164527
[332]	train-rmse:0.163831	eval-rmse:0.164288
[333]	train-rmse:0.16359	eval-rmse:0.164008
[334]	train-rmse:0.163367	eval-rmse:0.163699
[335]	train-rmse:0.163133	eval-rmse:0.163415
[336]	train-rmse:0.162833	eval-rmse:0.163104
[337]	train-rmse:0.162578	eval-rmse:0.162879
[338]	train-rmse:0.162447	eval-rmse:0.162746
[339]	train-rmse:0.162206	eval-rmse:0.162486
[340]	train-rmse:0.161988	eval-rmse:0.162203
[341]	train-rmse:0.161837	eval-rmse:0.162007
[342]	train-rmse:0.16166	eval-rmse:0.161807
[343]	train-rmse:0.161418	eval-rmse:0.161502
[344]	train-rmse:0.161222	eval-rmse:0.161321
[345]	train-rmse:0.16107	eval-rmse:0.161146
[346]	train-rmse:0.160776	eval-rmse:0.160854
[347]	train-rmse:0.160607	eval-rmse:0.160657
[348]	train-rmse:0.160397	eval-rmse:0.160369
[349]	train-rmse:0.160109	eval-rmse:0.160017
[350]	train-rmse:0.15985	eval-rmse:0.159681
[351]	train-rmse:0.159681	eval-rmse:0.159486
[352]	train-rmse:0.159438	eval-rmse:0.15911
[353]	train-rmse:0.159154	eval-rmse:0.158708
[354]	train-rmse:0.158934	eval-rmse:0.15845
[355]	train-rmse:0.158735	eval-rmse:0.158236
[356]	train-rmse:0.158474	eval-rmse:0.157874
[357]	train-rmse:0.158275	eval-rmse:0.157645
[358]	train-rmse:0.158049	eval-rmse:0.157418
[359]	train-rmse:0.157843	eval-rmse:0.157232
[360]	train-rmse:0.157664	eval-rmse:0.157026
[361]	train-rmse:0.157488	eval-rmse:0.156811
[362]	train-rmse:0.15728	eval-rmse:0.156613
[363]	train-rmse:0.157064	eval-rmse:0.156373
[364]	train-rmse:0.156831	eval-rmse:0.156136
[365]	train-rmse:0.156649	eval-rmse:0.155935
[366]	train-rmse:0.156479	eval-rmse:0.155737
[367]	train-rmse:0.156294	eval-rmse:0.155528
[368]	train-rmse:0.156066	eval-rmse:0.155179
[369]	train-rmse:0.155858	eval-rmse:0.154915
[370]	train-rmse:0.155605	eval-rmse:0.154658
[371]	train-rmse:0.155392	eval-rmse:0.154364
[372]	train-rmse:0.155214	eval-rmse:0.154156
[373]	train-rmse:0.154991	eval-rmse:0.153916
[374]	train-rmse:0.154794	eval-rmse:0.153693
[375]	train-rmse:0.154607	eval-rmse:0.153476
[376]	train-rmse:0.154451	eval-rmse:0.153326
[377]	train-rmse:0.154257	eval-rmse:0.153135
[378]	train-rmse:0.154011	eval-rmse:0.152844
[379]	train-rmse:0.153769	eval-rmse:0.152616
[380]	train-rmse:0.153571	eval-rmse:0.152468
[381]	train-rmse:0.153365	eval-rmse:0.152261
[382]	train-rmse:0.153188	eval-rmse:0.151991
[383]	train-rmse:0.153047	eval-rmse:0.151824
[384]	train-rmse:0.152912	eval-rmse:0.151676
[385]	train-rmse:0.152731	eval-rmse:0.151519
[386]	train-rmse:0.152538	eval-rmse:0.151273
[387]	train-rmse:0.152372	eval-rmse:0.151109
[388]	train-rmse:0.152256	eval-rmse:0.150997
[389]	train-rmse:0.152091	eval-rmse:0.150779
[390]	train-rmse:0.151871	eval-rmse:0.15056
[391]	train-rmse:0.151691	eval-rmse:0.150345
[392]	train-rmse:0.151489	eval-rmse:0.149996
[393]	train-rmse:0.151336	eval-rmse:0.149807
[394]	train-rmse:0.151202	eval-rmse:0.149678
[395]	train-rmse:0.150981	eval-rmse:0.14938
[396]	train-rmse:0.150718	eval-rmse:0.149044
[397]	train-rmse:0.150566	eval-rmse:0.148879
[398]	train-rmse:0.150394	eval-rmse:0.148657
[399]	train-rmse:0.150228	eval-rmse:0.148461
[400]	train-rmse:0.150131	eval-rmse:0.148315
[401]	train-rmse:0.149968	eval-rmse:0.148157
[402]	train-rmse:0.149775	eval-rmse:0.147935
[403]	train-rmse:0.149626	eval-rmse:0.14775
[404]	train-rmse:0.149504	eval-rmse:0.147572
[405]	train-rmse:0.149313	eval-rmse:0.14743
[406]	train-rmse:0.14915	eval-rmse:0.147247
[407]	train-rmse:0.148981	eval-rmse:0.147044
[408]	train-rmse:0.148801	eval-rmse:0.146836
[409]	train-rmse:0.148638	eval-rmse:0.14664
[410]	train-rmse:0.148482	eval-rmse:0.146484
[411]	train-rmse:0.148286	eval-rmse:0.146268
[412]	train-rmse:0.148149	eval-rmse:0.146109
[413]	train-rmse:0.147988	eval-rmse:0.145895
[414]	train-rmse:0.147817	eval-rmse:0.145679
[415]	train-rmse:0.147605	eval-rmse:0.145547
[416]	train-rmse:0.147463	eval-rmse:0.145402
[417]	train-rmse:0.147277	eval-rmse:0.145244
[418]	train-rmse:0.147105	eval-rmse:0.145054
[419]	train-rmse:0.146989	eval-rmse:0.144866
[420]	train-rmse:0.146794	eval-rmse:0.144706
[421]	train-rmse:0.146701	eval-rmse:0.144627
[422]	train-rmse:0.146477	eval-rmse:0.14436
[423]	train-rmse:0.146313	eval-rmse:0.144221
[424]	train-rmse:0.1462	eval-rmse:0.144085
[425]	train-rmse:0.146066	eval-rmse:0.143929
[426]	train-rmse:0.145882	eval-rmse:0.143765
[427]	train-rmse:0.145744	eval-rmse:0.143639
[428]	train-rmse:0.145639	eval-rmse:0.143595
[429]	train-rmse:0.145516	eval-rmse:0.143452
[430]	train-rmse:0.14531	eval-rmse:0.143175
[431]	train-rmse:0.145142	eval-rmse:0.142957
[432]	train-rmse:0.144966	eval-rmse:0.142784
[433]	train-rmse:0.144853	eval-rmse:0.142675
[434]	train-rmse:0.144798	eval-rmse:0.142609
[435]	train-rmse:0.14464	eval-rmse:0.142401
[436]	train-rmse:0.14451	eval-rmse:0.14226
[437]	train-rmse:0.144391	eval-rmse:0.142137
[438]	train-rmse:0.144201	eval-rmse:0.141939
[439]	train-rmse:0.144007	eval-rmse:0.141804
[440]	train-rmse:0.143864	eval-rmse:0.141624
[441]	train-rmse:0.143744	eval-rmse:0.141481
[442]	train-rmse:0.143591	eval-rmse:0.141358
[443]	train-rmse:0.143476	eval-rmse:0.141262
[444]	train-rmse:0.143322	eval-rmse:0.141109
[445]	train-rmse:0.143195	eval-rmse:0.140984
[446]	train-rmse:0.143047	eval-rmse:0.140853
[447]	train-rmse:0.142952	eval-rmse:0.140742
[448]	train-rmse:0.142756	eval-rmse:0.140562
[449]	train-rmse:0.142683	eval-rmse:0.140447
[450]	train-rmse:0.142557	eval-rmse:0.140325
[451]	train-rmse:0.14243	eval-rmse:0.140195
[452]	train-rmse:0.142261	eval-rmse:0.140004
[453]	train-rmse:0.142111	eval-rmse:0.139803
[454]	train-rmse:0.141881	eval-rmse:0.139646
[455]	train-rmse:0.141826	eval-rmse:0.139634
[456]	train-rmse:0.141733	eval-rmse:0.139536
[457]	train-rmse:0.14164	eval-rmse:0.139422
[458]	train-rmse:0.141556	eval-rmse:0.139311
[459]	train-rmse:0.141471	eval-rmse:0.139167
[460]	train-rmse:0.141333	eval-rmse:0.138978
[461]	train-rmse:0.141208	eval-rmse:0.138892
[462]	train-rmse:0.141105	eval-rmse:0.138723
[463]	train-rmse:0.141	eval-rmse:0.138614
[464]	train-rmse:0.140925	eval-rmse:0.13854
[465]	train-rmse:0.140808	eval-rmse:0.138414
[466]	train-rmse:0.140659	eval-rmse:0.138241
[467]	train-rmse:0.140581	eval-rmse:0.138138
[468]	train-rmse:0.140456	eval-rmse:0.138026
[469]	train-rmse:0.140302	eval-rmse:0.137855
[470]	train-rmse:0.140268	eval-rmse:0.137789
[471]	train-rmse:0.140139	eval-rmse:0.137658
[472]	train-rmse:0.140028	eval-rmse:0.137502
[473]	train-rmse:0.139964	eval-rmse:0.137423
[474]	train-rmse:0.13974	eval-rmse:0.137208
[475]	train-rmse:0.139642	eval-rmse:0.137097
[476]	train-rmse:0.139547	eval-rmse:0.136999
[477]	train-rmse:0.139401	eval-rmse:0.136779
[478]	train-rmse:0.139243	eval-rmse:0.136595
[479]	train-rmse:0.139154	eval-rmse:0.136534
[480]	train-rmse:0.139077	eval-rmse:0.136447
[481]	train-rmse:0.138903	eval-rmse:0.136302
[482]	train-rmse:0.138778	eval-rmse:0.136166
[483]	train-rmse:0.138677	eval-rmse:0.136073
[484]	train-rmse:0.138499	eval-rmse:0.135908
[485]	train-rmse:0.13844	eval-rmse:0.135864
[486]	train-rmse:0.138372	eval-rmse:0.135812
[487]	train-rmse:0.138238	eval-rmse:0.135707
[488]	train-rmse:0.138142	eval-rmse:0.135636
[489]	train-rmse:0.138062	eval-rmse:0.135539
[490]	train-rmse:0.137959	eval-rmse:0.135437
[491]	train-rmse:0.137842	eval-rmse:0.135278
[492]	train-rmse:0.137752	eval-rmse:0.135162
[493]	train-rmse:0.13763	eval-rmse:0.135039
[494]	train-rmse:0.137496	eval-rmse:0.134938
[495]	train-rmse:0.137383	eval-rmse:0.134751
[496]	train-rmse:0.137294	eval-rmse:0.134767
[497]	train-rmse:0.137201	eval-rmse:0.134714
[498]	train-rmse:0.137112	eval-rmse:0.134662
[499]	train-rmse:0.137042	eval-rmse:0.134608
[500]	train-rmse:0.136936	eval-rmse:0.134494
[501]	train-rmse:0.136829	eval-rmse:0.13441
[502]	train-rmse:0.136721	eval-rmse:0.134388
[503]	train-rmse:0.136617	eval-rmse:0.134281
[504]	train-rmse:0.136537	eval-rmse:0.134204
[505]	train-rmse:0.136442	eval-rmse:0.134118
[506]	train-rmse:0.136328	eval-rmse:0.134026
[507]	train-rmse:0.136283	eval-rmse:0.134006
[508]	train-rmse:0.136236	eval-rmse:0.133974
[509]	train-rmse:0.136128	eval-rmse:0.133824
[510]	train-rmse:0.136039	eval-rmse:0.133748
[511]	train-rmse:0.135982	eval-rmse:0.133682
[512]	train-rmse:0.135946	eval-rmse:0.133637
[513]	train-rmse:0.135872	eval-rmse:0.133543
[514]	train-rmse:0.135791	eval-rmse:0.13345
[515]	train-rmse:0.135724	eval-rmse:0.133427
[516]	train-rmse:0.135664	eval-rmse:0.133392
[517]	train-rmse:0.135633	eval-rmse:0.13336
[518]	train-rmse:0.135557	eval-rmse:0.133304
[519]	train-rmse:0.13548	eval-rmse:0.133242
[520]	train-rmse:0.135435	eval-rmse:0.133236
[521]	train-rmse:0.135412	eval-rmse:0.133198
[522]	train-rmse:0.135357	eval-rmse:0.13327
[523]	train-rmse:0.135278	eval-rmse:0.133177
[524]	train-rmse:0.1352	eval-rmse:0.133176
[525]	train-rmse:0.13515	eval-rmse:0.133152
[526]	train-rmse:0.135064	eval-rmse:0.133027
[527]	train-rmse:0.135008	eval-rmse:0.132949
[528]	train-rmse:0.134937	eval-rmse:0.132853
[529]	train-rmse:0.134825	eval-rmse:0.132765
[530]	train-rmse:0.134749	eval-rmse:0.132718
[531]	train-rmse:0.134694	eval-rmse:0.132664
[532]	train-rmse:0.134569	eval-rmse:0.132543
[533]	train-rmse:0.134457	eval-rmse:0.132402
[534]	train-rmse:0.134356	eval-rmse:0.132294
[535]	train-rmse:0.134261	eval-rmse:0.132259
[536]	train-rmse:0.134173	eval-rmse:0.132232
[537]	train-rmse:0.134118	eval-rmse:0.132198
[538]	train-rmse:0.134042	eval-rmse:0.132123
[539]	train-rmse:0.133966	eval-rmse:0.132089
[540]	train-rmse:0.133888	eval-rmse:0.132023
[541]	train-rmse:0.133848	eval-rmse:0.131931
[542]	train-rmse:0.133798	eval-rmse:0.131853
[543]	train-rmse:0.133732	eval-rmse:0.131796
[544]	train-rmse:0.133714	eval-rmse:0.131772
[545]	train-rmse:0.13365	eval-rmse:0.131721
[546]	train-rmse:0.133571	eval-rmse:0.131642
[547]	train-rmse:0.133493	eval-rmse:0.131581
[548]	train-rmse:0.133409	eval-rmse:0.131504
[549]	train-rmse:0.133372	eval-rmse:0.131453
[550]	train-rmse:0.133294	eval-rmse:0.13139
[551]	train-rmse:0.133213	eval-rmse:0.131307
[552]	train-rmse:0.133161	eval-rmse:0.131212
[553]	train-rmse:0.133084	eval-rmse:0.131121
[554]	train-rmse:0.133002	eval-rmse:0.131035
[555]	train-rmse:0.132957	eval-rmse:0.131009
[556]	train-rmse:0.132899	eval-rmse:0.130947
[557]	train-rmse:0.132824	eval-rmse:0.130891
[558]	train-rmse:0.132743	eval-rmse:0.130756
[559]	train-rmse:0.132666	eval-rmse:0.130715
[560]	train-rmse:0.132608	eval-rmse:0.130645
[561]	train-rmse:0.132564	eval-rmse:0.130556
[562]	train-rmse:0.132518	eval-rmse:0.130528
[563]	train-rmse:0.132473	eval-rmse:0.130462
[564]	train-rmse:0.13241	eval-rmse:0.130401
[565]	train-rmse:0.132367	eval-rmse:0.130295
[566]	train-rmse:0.132298	eval-rmse:0.130249
[567]	train-rmse:0.132252	eval-rmse:0.13019
[568]	train-rmse:0.132208	eval-rmse:0.130173
[569]	train-rmse:0.132163	eval-rmse:0.130095
[570]	train-rmse:0.132095	eval-rmse:0.13007
[571]	train-rmse:0.132026	eval-rmse:0.130016
[572]	train-rmse:0.131994	eval-rmse:0.129973
[573]	train-rmse:0.131936	eval-rmse:0.129909
[574]	train-rmse:0.131894	eval-rmse:0.129874
[575]	train-rmse:0.131807	eval-rmse:0.129826
[576]	train-rmse:0.131764	eval-rmse:0.129829
[577]	train-rmse:0.131717	eval-rmse:0.129797
[578]	train-rmse:0.131634	eval-rmse:0.129687
[579]	train-rmse:0.131569	eval-rmse:0.129628
[580]	train-rmse:0.131499	eval-rmse:0.129587
[581]	train-rmse:0.13144	eval-rmse:0.12953
[582]	train-rmse:0.131373	eval-rmse:0.129445
[583]	train-rmse:0.131319	eval-rmse:0.129521
[584]	train-rmse:0.131285	eval-rmse:0.129478
[585]	train-rmse:0.131226	eval-rmse:0.129535
[586]	train-rmse:0.131181	eval-rmse:0.129483
[587]	train-rmse:0.131121	eval-rmse:0.129461
[588]	train-rmse:0.13107	eval-rmse:0.129397
[589]	train-rmse:0.131016	eval-rmse:0.129338
[590]	train-rmse:0.130979	eval-rmse:0.129289
[591]	train-rmse:0.130928	eval-rmse:0.129258
[592]	train-rmse:0.130892	eval-rmse:0.129269
[593]	train-rmse:0.130867	eval-rmse:0.129196
[594]	train-rmse:0.130839	eval-rmse:0.129169
[595]	train-rmse:0.130815	eval-rmse:0.129138
[596]	train-rmse:0.130775	eval-rmse:0.129102
[597]	train-rmse:0.130755	eval-rmse:0.129136
[598]	train-rmse:0.130732	eval-rmse:0.129105
[599]	train-rmse:0.130681	eval-rmse:0.129074
[600]	train-rmse:0.130635	eval-rmse:0.129027
[601]	train-rmse:0.130597	eval-rmse:0.129008
[602]	train-rmse:0.130599	eval-rmse:0.128936
[603]	train-rmse:0.130559	eval-rmse:0.128924
[604]	train-rmse:0.13054	eval-rmse:0.128921
[605]	train-rmse:0.130514	eval-rmse:0.128905
[606]	train-rmse:0.130473	eval-rmse:0.128888
[607]	train-rmse:0.130415	eval-rmse:0.128885
[608]	train-rmse:0.130377	eval-rmse:0.128834
[609]	train-rmse:0.130337	eval-rmse:0.128791
[610]	train-rmse:0.130314	eval-rmse:0.128775
[611]	train-rmse:0.130303	eval-rmse:0.128767
[612]	train-rmse:0.130256	eval-rmse:0.128678
[613]	train-rmse:0.130236	eval-rmse:0.128645
[614]	train-rmse:0.130202	eval-rmse:0.128623
[615]	train-rmse:0.130143	eval-rmse:0.128765
[616]	train-rmse:0.130115	eval-rmse:0.128728
[617]	train-rmse:0.130074	eval-rmse:0.128698
[618]	train-rmse:0.130041	eval-rmse:0.128687
[619]	train-rmse:0.130008	eval-rmse:0.128623
[620]	train-rmse:0.129996	eval-rmse:0.128626
[621]	train-rmse:0.129944	eval-rmse:0.128559
[622]	train-rmse:0.129922	eval-rmse:0.128529
[623]	train-rmse:0.129914	eval-rmse:0.128674
[624]	train-rmse:0.129852	eval-rmse:0.128786
[625]	train-rmse:0.129807	eval-rmse:0.12879
[626]	train-rmse:0.129769	eval-rmse:0.128772
[627]	train-rmse:0.12977	eval-rmse:0.128738
[628]	train-rmse:0.129712	eval-rmse:0.128763
[629]	train-rmse:0.129662	eval-rmse:0.128724
[630]	train-rmse:0.129646	eval-rmse:0.128697
[631]	train-rmse:0.129638	eval-rmse:0.128685
[632]	train-rmse:0.129613	eval-rmse:0.128704
[633]	train-rmse:0.129574	eval-rmse:0.128689
[634]	train-rmse:0.129563	eval-rmse:0.128682
[635]	train-rmse:0.129542	eval-rmse:0.128676
[636]	train-rmse:0.129525	eval-rmse:0.128666
[637]	train-rmse:0.129498	eval-rmse:0.128644
[638]	train-rmse:0.129471	eval-rmse:0.128598
[639]	train-rmse:0.129443	eval-rmse:0.128546
[640]	train-rmse:0.129447	eval-rmse:0.128526
[641]	train-rmse:0.129423	eval-rmse:0.128514
[642]	train-rmse:0.129409	eval-rmse:0.12843
[643]	train-rmse:0.129375	eval-rmse:0.128396
[644]	train-rmse:0.129351	eval-rmse:0.128385
[645]	train-rmse:0.129324	eval-rmse:0.128344
[646]	train-rmse:0.129305	eval-rmse:0.128311
[647]	train-rmse:0.129311	eval-rmse:0.128371
[648]	train-rmse:0.129296	eval-rmse:0.12839
[649]	train-rmse:0.129274	eval-rmse:0.128388
[650]	train-rmse:0.129198	eval-rmse:0.12839
[651]	train-rmse:0.129169	eval-rmse:0.128392
[652]	train-rmse:0.129138	eval-rmse:0.128373
[653]	train-rmse:0.129122	eval-rmse:0.128363
[654]	train-rmse:0.129106	eval-rmse:0.128338
[655]	train-rmse:0.129088	eval-rmse:0.128335
[656]	train-rmse:0.129077	eval-rmse:0.128329
[657]	train-rmse:0.129061	eval-rmse:0.128335
[658]	train-rmse:0.129046	eval-rmse:0.128396
[659]	train-rmse:0.129042	eval-rmse:0.1284
[660]	train-rmse:0.129036	eval-rmse:0.128382
[661]	train-rmse:0.128992	eval-rmse:0.128364
[662]	train-rmse:0.128975	eval-rmse:0.128357
[663]	train-rmse:0.128968	eval-rmse:0.128432
[664]	train-rmse:0.128942	eval-rmse:0.128374
[665]	train-rmse:0.128907	eval-rmse:0.12835
[666]	train-rmse:0.12888	eval-rmse:0.128407
[667]	train-rmse:0.128839	eval-rmse:0.128392
[668]	train-rmse:0.128809	eval-rmse:0.128371
[669]	train-rmse:0.128787	eval-rmse:0.128366
[670]	train-rmse:0.128771	eval-rmse:0.128343
[671]	train-rmse:0.128772	eval-rmse:0.128314
[672]	train-rmse:0.128753	eval-rmse:0.128321
[673]	train-rmse:0.128724	eval-rmse:0.128291
[674]	train-rmse:0.12873	eval-rmse:0.128288
[675]	train-rmse:0.128672	eval-rmse:0.128273
[676]	train-rmse:0.128654	eval-rmse:0.12825
[677]	train-rmse:0.128644	eval-rmse:0.128257
[678]	train-rmse:0.128622	eval-rmse:0.128256
[679]	train-rmse:0.128604	eval-rmse:0.128277
[680]	train-rmse:0.12861	eval-rmse:0.128363
[681]	train-rmse:0.128581	eval-rmse:0.128333
[682]	train-rmse:0.128568	eval-rmse:0.128346
[683]	train-rmse:0.128569	eval-rmse:0.128332
[684]	train-rmse:0.128572	eval-rmse:0.128371
[685]	train-rmse:0.128524	eval-rmse:0.128304
[686]	train-rmse:0.128494	eval-rmse:0.128296
[687]	train-rmse:0.128476	eval-rmse:0.128289
[688]	train-rmse:0.128437	eval-rmse:0.12829
[689]	train-rmse:0.128452	eval-rmse:0.128377
[690]	train-rmse:0.128412	eval-rmse:0.128444
[691]	train-rmse:0.128379	eval-rmse:0.128408
[692]	train-rmse:0.128377	eval-rmse:0.128449
[693]	train-rmse:0.128345	eval-rmse:0.128448
[694]	train-rmse:0.128316	eval-rmse:0.128414
[695]	train-rmse:0.128302	eval-rmse:0.128396
[696]	train-rmse:0.128267	eval-rmse:0.128439
[697]	train-rmse:0.128233	eval-rmse:0.128498
[698]	train-rmse:0.12825	eval-rmse:0.128482
[699]	train-rmse:0.128203	eval-rmse:0.128429
[700]	train-rmse:0.1282	eval-rmse:0.128433
[701]	train-rmse:0.128185	eval-rmse:0.128468
[702]	train-rmse:0.128156	eval-rmse:0.128457
[703]	train-rmse:0.128131	eval-rmse:0.128482
[704]	train-rmse:0.128111	eval-rmse:0.128505
[705]	train-rmse:0.128101	eval-rmse:0.12853
[706]	train-rmse:0.128093	eval-rmse:0.128544
[707]	train-rmse:0.128066	eval-rmse:0.128551
[708]	train-rmse:0.128064	eval-rmse:0.128559
[709]	train-rmse:0.128027	eval-rmse:0.128572
[710]	train-rmse:0.128	eval-rmse:0.128569
[711]	train-rmse:0.127968	eval-rmse:0.128541
[712]	train-rmse:0.127961	eval-rmse:0.128529
[713]	train-rmse:0.127967	eval-rmse:0.128513
[714]	train-rmse:0.127971	eval-rmse:0.12853
[715]	train-rmse:0.127981	eval-rmse:0.128579
[716]	train-rmse:0.127963	eval-rmse:0.128577
[717]	train-rmse:0.12791	eval-rmse:0.128532
[718]	train-rmse:0.127894	eval-rmse:0.128544
[719]	train-rmse:0.127882	eval-rmse:0.128576
[720]	train-rmse:0.127869	eval-rmse:0.128561
[721]	train-rmse:0.127884	eval-rmse:0.128586
[722]	train-rmse:0.127902	eval-rmse:0.128597
[723]	train-rmse:0.127907	eval-rmse:0.128563
[724]	train-rmse:0.127915	eval-rmse:0.128552
[725]	train-rmse:0.127913	eval-rmse:0.128525
[726]	train-rmse:0.127869	eval-rmse:0.128499
Stopping. Best iteration:
[676]	train-rmse:0.128654	eval-rmse:0.12825

In [10]:

    

_ = xgb.plot_importance(xgb_model1)


    

In [11]:

    

def par_dep(xs, frame, model, resolution=20, bins=None):
    
    """ Creates Pandas dataframe containing partial dependence for a single variable.
    
    Args:
        xs: Variable for which to calculate partial dependence.
        frame: H2OFrame for which to calculate partial dependence.
        model: XGBoost model for which to calculate partial dependence.
        resolution: The number of points across the domain of xs for which to calculate partial dependence.
    
    Returns:
        Pandas dataframe containing partial dependence values.
        
    """
    
    # don't show progress bars for parse
    h2o.no_progress()
    
    # init empty Pandas frame w/ correct col names
    par_dep_frame = pd.DataFrame(columns=[xs, 'partial_dependence'])
    
    # cache original data 
    col_cache = h2o.deep_copy(frame[xs], xid='col_cache')
  
    # determine values at which to calculate partial dependency
    if bins == None:
        min_ = frame[xs].min()
        max_ = frame[xs].max()
        by = (max_ - min_)/resolution
        bins = np.arange(min_, max_, by)
        
    # calculate partial dependency  
    # by setting column of interest to constant 
    for j in bins:
        frame[xs] = j
        dframe = xgb.DMatrix(frame.as_data_frame(),)
        par_dep_i = h2o.H2OFrame(model.predict(dframe).tolist())
        par_dep_j = par_dep_i.mean()[0]
        par_dep_frame = par_dep_frame.append({xs:j,
                                              'partial_dependence': par_dep_j}, 
                                              ignore_index=True)
        
    # return input frame to original cached state    
    frame[xs] = h2o.get_frame('col_cache')

    return par_dep_frame


    

In [12]:

    

par_dep_OverallCond = par_dep('OverallCond', valid[reals], xgb_model1)
par_dep_GrLivArea = par_dep('GrLivArea', valid[reals], xgb_model1)
par_dep_LotArea = par_dep('LotArea', valid[reals], xgb_model1)


    

In [13]:

    

def get_quantile_dict(y, id_, frame):

    """ Returns the percentiles of a column y as the indices for another column id_.
    
    Args:
        y: Column in which to find percentiles.
        id_: Id column that stores indices for percentiles of y.
        frame: H2OFrame containing y and id_. 
    
    Returns:
        Dictionary of percentile values and index column values.
    
    """
    
    quantiles_df = frame.as_data_frame()
    quantiles_df.sort_values(y, inplace=True)
    quantiles_df.reset_index(inplace=True)
    
    percentiles_dict = {}
    percentiles_dict[0] = quantiles_df.loc[0, id_]
    percentiles_dict[99] = quantiles_df.loc[quantiles_df.shape[0]-1, id_]
    inc = quantiles_df.shape[0]//10
    
    for i in range(1, 10):
        percentiles_dict[i * 10] = quantiles_df.loc[i * inc,  id_]

    return percentiles_dict


    

In [14]:

    

quantile_dict = get_quantile_dict('SalePrice', 'Id', valid)


    

In [15]:

    

bins_OverallCond = list(par_dep_OverallCond['OverallCond'])
bins_GrLivArea = list(par_dep_GrLivArea['GrLivArea'])
bins_LotArea = list(par_dep_LotArea['LotArea'])

for i in sorted(quantile_dict.keys()):
    
    col_name = 'Percentile_' + str(i)
    
    par_dep_OverallCond[col_name] = par_dep('OverallCond', 
                                            valid[valid['Id'] == int(quantile_dict[i])][reals], 
                                            xgb_model1, 
                                            bins=bins_OverallCond)['partial_dependence']
    
    par_dep_GrLivArea[col_name] = par_dep('GrLivArea', 
                                          valid[valid['Id'] == int(quantile_dict[i])][reals], 
                                          xgb_model1, 
                                          bins=bins_GrLivArea)['partial_dependence']
    
    par_dep_LotArea[col_name] = par_dep('LotArea', 
                                         valid[valid['Id'] == int(quantile_dict[i])][reals], 
                                         xgb_model1, 
                                         bins=bins_LotArea)['partial_dependence']


    

In [16]:

    

# OverallCond

fig, ax = plt.subplots()
par_dep_OverallCond.drop('partial_dependence', axis=1).plot(x='OverallCond', colormap='gnuplot', ax=ax)

par_dep_OverallCond.plot(title='Partial Dependence and ICE for OverallCond',
                         x='OverallCond', 
                         y='partial_dependence',
                         style='r-', 
                         linewidth=3, 
                         ax=ax)

_ = plt.legend(bbox_to_anchor=(1.05, 0),
               loc=3, 
               borderaxespad=0.)


    

In [17]:

    

# GrLivArea

fig, ax = plt.subplots()
par_dep_GrLivArea.drop('partial_dependence', axis=1).plot(x='GrLivArea', colormap='gnuplot', ax=ax)

par_dep_GrLivArea.plot(title='Partial Dependence and ICE for GrLivArea',
                         x='GrLivArea', 
                         y='partial_dependence',
                         style='r-', 
                         linewidth=3, 
                         ax=ax)

_ = plt.legend(bbox_to_anchor=(1.05, 0),
               loc=3, 
               borderaxespad=0.)


    

In [18]:

    

# LotArea

fig, ax = plt.subplots()
par_dep_LotArea.drop('partial_dependence', axis=1).plot(x='LotArea', colormap='gnuplot', ax=ax)

par_dep_LotArea.plot(title='Partial Dependence and ICE for LotArea',
                     x='LotArea', 
                     y='partial_dependence',
                     style='r-', 
                     linewidth=3, 
                     ax=ax)

_ = plt.legend(bbox_to_anchor=(1.05, 0),
               loc=3, 
               borderaxespad=0.)


    

In [19]:

    

h2o.cluster().shutdown(prompt=True)


    

    




Are you sure you want to shutdown the H2O instance running at http://127.0.0.1:54321 (Y/N)? y
H2O session _sid_91ac closed.