R H2O - Loss vs Epoch

Question

I'm working on ANN model (a binary classification problem) in R using the h2o package because neuralnet takes too long with the data set I'm working on .

model <- h2o.deeplearning(model_id = "h2o_ANN" ,
                          training_frame = aa.h2o.tra ,
                          seed = 1 ,y = target ,
                          x = predictors ,epochs = 500 ,
                          hidden = 16 ,adaptive_rate = F ,
                          rate = 0.01 ,activation = "Tanh")

However, when I plot loss vs epochs, the line graph looks weird (for me!). I only have a small number of values within the linechart. Perhaps this is because h2o stores only a summary of the results

loss vs epoch

How can I plot the results for each epoch?

This is a sample of the data I am using.

       LIMIT_BAL        SEX  EDUCATION   MARRIAGE        AGE
6756   0.7065360 -0.8112315 -1.2947658 -0.9235809 -0.2603721
8362  -1.0616096 -0.8112315  0.5637609  1.0827051  1.2644609
25951 -0.9847337  1.2326514 -1.2947658 -0.9235809 -1.0227887
18182 -0.5234783 -0.8112315  0.5637609  1.0827051 -0.6960387
28379  1.7827986  1.2326514  0.5637609 -0.9235809 -1.0227887
23527  0.2452807  1.2326514  0.5637609 -0.9235809 -0.8049554
           PAY_1      PAY_2      PAY_3      PAY_4      PAY_5
6756  0.01322092  0.1083065  0.1357685  0.1858918  0.2313689
8362  0.01322092  0.1083065  0.1357685  0.1858918  0.2313689
25951 0.90161187 -1.5567129 -1.5299515 -1.5196811 -0.6477726
18182 0.01322092  0.1083065  0.1357685  0.1858918  0.2313689
28379 0.90161187 -0.7242032 -0.6970915 -0.6668946 -0.6477726
23527 0.01322092  0.1083065  0.1357685 -0.6668946 -0.6477726
           PAY_6  BILL_AMT1  BILL_AMT2  BILL_AMT3  BILL_AMT4
6756   0.2495171 -0.4432728 -0.4042181 -0.3504430 -0.3479263
8362   0.2495171 -0.2973090 -0.2785499 -0.2412887  0.2332462
25951  0.2495171 -0.6944606 -0.6900911 -0.6772422 -0.6719405
18182  0.2495171  0.9112907  0.9514314  0.9672028  1.0530965
28379 -0.6177853 -0.6816971 -0.6851955 -0.6730279 -0.4750689
23527 -0.6177853  0.6555853  0.7186927 -0.1841930 -0.6608832
       BILL_AMT5  BILL_AMT6     PAY_AMT1    PAY_AMT2
6756  -0.2428634 -0.1717688 -0.166281672 -0.12988389
8362  -0.1785104 -0.1490618 -0.260590656 -0.19971668
25951 -0.6206216 -0.5939683 -0.355530470 -0.26565731
18182  1.0982856  1.1123754 -0.001382886 -0.01221359
28379 -0.6363075 -0.6281210 -0.333451444 -0.25244203
23527 -0.6517141 -0.6390807 -0.021190927 -0.12988389
         PAY_AMT3    PAY_AMT4    PAY_AMT5    PAY_AMT6 DEFAULT
6756  -0.18856887  0.05263758 -0.06912059 -0.23740829       0
8362  -0.24714368 -0.23512972 -0.21064007 -0.23740829       0
25951 -0.30762337 -0.14425583 -0.25017065 -0.18093455       0
18182  0.01977649 -0.02107557  0.01336656  0.04496042       0
28379  0.44700354 -0.20633349 -0.21617435  0.38476293       0
23527  0.21097801 -0.26730609 -0.25926268 -0.25412452       0

Hi Lirios, welcome to Stack Overflow. It will be much easier to help if you provide at least a sample of your data with `dput(aa.h2o.tra[1:20,])`. You can edit your question and paste the output. You can surround it with three backticks (```) for better formatting. See [How to make a reproducible example](https://stackoverflow.com/questions/5963269/) for more info. — Ian Campbell, May 31 '20 at 13:30

R H2O - Loss vs Epoch

0 Answers0