I'm new to H2O project but have a lot of interest to use it for my web application. I've made a few tutorials building a model with h2o
library in R and then exporting it as POJO
model.
My problem is that the data I need to predict is in different format I used for train dataset. I am wondering is it possible to use my R code with data transformation and export it as POJO
(or similar) so I could prepare a new data for prediction in real time.
I've found some examples in "Productionizing H2O" section on H2O docs page (link), like Consumer loan application
and Storm bolt
. But in first case they use the same set of features and in the second one data prep is done with .java code (not sure on that, I have a poor programming skills outside of R). So I am not able to figure out how to use my transformation code in R on a new dataset.
Any suggestions?