In [1]:

    

import h2o
from h2o.estimators.deeplearning import H2ODeepLearningEstimator


    

In [2]:

    

h2o.init()


    

    




Warning: Version mismatch. H2O is version 3.5.0.99999, but the python package is version UNKNOWN.






    





H2O cluster uptime: 
58 minutes 48 seconds 43 milliseconds 
H2O cluster version: 
3.5.0.99999
H2O cluster name: 
ludirehak
H2O cluster total nodes: 
1
H2O cluster total memory: 
4.44 GB
H2O cluster total cores: 
8
H2O cluster allowed cores: 
8
H2O cluster healthy: 
True
H2O Connection ip: 
127.0.0.1
H2O Connection port: 
54321

In [3]:

    

from h2o.utils.shared_utils import _locate # private function. used to find files within h2o git project directory.

prostate = h2o.upload_file(path=_locate("smalldata/logreg/prostate.csv"))
prostate.describe()


    

    




Parse Progress: [##################################################] 100%
Uploaded py2a71800e-2ec6-4f71-b955-854a4f22aeb3 into cluster with 380 rows and 9 cols
Rows: 380 Cols: 9

Chunk compression summary:






    





chunk_type
chunk_name
count
count_percentage
size
size_percentage
CBS
Bits
1
11.111112
    118  B
2.4210093
C1N
1-Byte Integers (w/o NAs)
5
55.555557
    2.2 KB
45.958145
C2
2-Byte Integers
1
11.111112
    828  B
16.9881
C2S
2-Byte Fractions
2
22.222223
    1.6 KB
34.632744






    




Frame distribution summary:






    






size
number_of_rows
number_of_chunks_per_column
number_of_chunks
172.16.2.37:54321
    4.8 KB
380.0
1.0
9.0
mean
    4.8 KB
380.0
1.0
9.0
min
    4.8 KB
380.0
1.0
9.0
max
    4.8 KB
380.0
1.0
9.0
stddev
      0  B
0.0
0.0
0.0
total
    4.8 KB
380.0
1.0
9.0






    




Column-by-Column Summary:







    






ID
CAPSULE
AGE
RACE
DPROS
DCAPS
PSA
VOL
GLEASON
type
int
int
int
int
int
int
real
real
int
mins
1.0
0.0
43.0
0.0
1.0
1.0
0.3
0.0
0.0
maxs
380.0
1.0
79.0
2.0
4.0
2.0
139.7
97.6
9.0
mean
190.5
0.4
66.0
1.1
2.3
1.1
15.4
15.8
6.4
sigma
109.8
0.5
6.5
0.3
1.0
0.3
20.0
18.3
1.1
zero_count
0
227
0
3
0
0
0
167
2
missing_count
0
0
0
0
0
0
0
0
0

In [4]:

    

prostate["CAPSULE"] = prostate["CAPSULE"].asfactor()
model = H2ODeepLearningEstimator(activation = "Tanh", hidden = [10, 10, 10], epochs = 10000)
model.train(x = list(set(prostate.columns) - set(["ID","CAPSULE"])), y ="CAPSULE", training_frame = prostate)
model.show()


    

    




deeplearning Model Build Progress: [##################################################] 100%
Model Details
=============
H2ODeepLearningEstimator :  Deep Learning
Model Key:  DeepLearning_model_python_1445544453075_137

Status of Neuron Layers: predicting CAPSULE, 2-class classification, bernoulli distribution, CrossEntropy loss, 322 weights/biases, 8.5 KB, 3,800,000 training samples, mini-batch size 1







    






layer
units
type
dropout
l1
l2
mean_rate
rate_RMS
momentum
mean_weight
weight_RMS
mean_bias
bias_RMS

1
7
Input
0.0










2
10
Tanh
0.0
0.0
0.0
0.1
0.1
0.0
0.1
1.0
0.4
0.7

3
10
Tanh
0.0
0.0
0.0
0.1
0.1
0.0
0.0
1.4
1.1
0.7

4
10
Tanh
0.0
0.0
0.0
0.2
0.2
0.0
-0.2
1.8
-0.2
0.8

5
2
Softmax

0.0
0.0
0.4
0.1
0.0
-0.2
5.7
0.1
0.3






    





ModelMetricsBinomial: deeplearning
** Reported on train data. **

MSE: 0.010708193224
R^2: 0.955478877615
LogLoss: 0.0689458344205
AUC: 0.996818404307
Gini: 0.993636808615

Confusion Matrix (Act/Pred) for max f1 @ threshold = 0.85259659057:






    






0
1
Error
Rate
0
224.0
3.0
0.0132
 (3.0/227.0)
1
0.0
153.0
0.0
 (0.0/153.0)
Total
224.0
156.0
0.0079
 (3.0/380.0)






    




Maximum Metrics: Maximum metrics at their respective thresholds







    





metric
threshold
value
idx
max f1
0.9
1.0
107.0
max f2
0.9
1.0
107.0
max f0point5
0.9
1.0
107.0
max accuracy
0.9
1.0
107.0
max precision
1.0
1.0
0.0
max absolute_MCC
0.9
1.0
107.0
max min_per_class_accuracy
0.9
1.0
105.0






    




Scoring History:






    






timestamp
duration
training_speed
epochs
samples
training_MSE
training_r2
training_logloss
training_AUC
training_classification_error

2015-10-22 14:06:21
 0.000 sec
None
0.0
0.0
nan
nan
nan
nan
nan

2015-10-22 14:06:21
 0.038 sec
115151 rows/sec
10.0
3800.0
0.2
0.2
0.6
0.8
0.3

2015-10-22 14:06:26
 5.047 sec
309126 rows/sec
4100.0
1558000.0
0.0
0.9
0.1
1.0
0.0

2015-10-22 14:06:31
10.051 sec
308783 rows/sec
8160.0
3100800.0
0.0
0.9
0.1
1.0
0.0

2015-10-22 14:06:34
12.360 sec
307717 rows/sec
10000.0
3800000.0
0.0
1.0
0.1
1.0
0.0

In [5]:

    

predictions = model.predict(prostate)
predictions.show()


    

    




H2OFrame with 380 rows and 3 columns: 






    







  

    

      

      
predict
      
p0
      
p1
    
  
  

    

      
0
      
0
      
9.993875e-01
      
6.125394e-04
    
    

      
1
      
0
      
9.999998e-01
      
1.937478e-07
    
    

      
2
      
0
      
9.999646e-01
      
3.535732e-05
    
    

      
3
      
0
      
1.000000e+00
      
2.235483e-12
    
    

      
4
      
0
      
9.999950e-01
      
5.024862e-06
    
    

      
5
      
1
      
1.237468e-07
      
9.999999e-01
    
    

      
6
      
0
      
9.992793e-01
      
7.206910e-04
    
    

      
7
      
0
      
1.000000e+00
      
9.146884e-19
    
    

      
8
      
0
      
1.000000e+00
      
8.434714e-13
    
    

      
9
      
0
      
9.999994e-01
      
6.112821e-07
    
  

In [6]:

    

performance = model.model_performance(prostate)
performance.show()


    

    




ModelMetricsBinomial: deeplearning
** Reported on test data. **

MSE: 0.010708193224
R^2: 0.955478877615
LogLoss: 0.0689458344205
AUC: 0.996804007947
Gini: 0.993608015894

Confusion Matrix (Act/Pred) for max f1 @ threshold = 0.85259659057:






    






0
1
Error
Rate
0
224.0
3.0
0.0132
 (3.0/227.0)
1
0.0
153.0
0.0
 (0.0/153.0)
Total
224.0
156.0
0.0079
 (3.0/380.0)






    




Maximum Metrics: Maximum metrics at their respective thresholds







    





metric
threshold
value
idx
max f1
0.9
1.0
155.0
max f2
0.9
1.0
155.0
max f0point5
0.9
1.0
155.0
max accuracy
0.9
1.0
155.0
max precision
1.0
1.0
0.0
max absolute_MCC
0.9
1.0
155.0
max min_per_class_accuracy
0.9
1.0
153.0