I am using Spark MLLib to make prediction and I would like to know if it is possible to create your custom Estimators.
More precisely, I have 4 variable, two categorical (country and gender) and two continuous (age and height) and I fitted a different model for each country, gender pair which make prediction given age and height. For example if I have three different countries in my training set: France, Spain and UK I have six models (one for (France, M), (France, F), (Spain, M), …
Would it be possible to combine all the steps I described previously in a single estimator (via Pipeline maybe ?)
Thanks in advance for your help