Auto Model only take 500000 records for training? Missing ID for scoring output?

tuan_nguyentuan_nguyen Member Posts: 1 Learner I
edited December 2019 in Help
I tried Auto Model and Deployment tab in a paid RM Studio Large having 16GB RAM (We don't have RM Server) and encounter some difficulties.

1) Auto Model -  When opening process after Auto Model ran, I noticed Auto Model only take 500k records for training while my input data set have around 30 millions records but only several fields (we probably have close to 100M records but 30M ~ 900MB for starter). Is there anyway we can do so that Auto Model train with more data?
I've tried to open up the process and increase the size of sampling and it can run fine on 10M. I can only imagine data grow more and more and we eventually add more features in the mix. But I want to use Auto Model as much as possible if it serve the purpose in later stage like Deployment

2) Deployment - I had indicated ID-field in data to be scored but after the output was generated, the result didn't have ID field that I expect. Any explanation and how we can export the result with ID-fields and their predictions so that we can import to our datawarehouse ?

Answers

  • sgenzersgenzer Administrator, Moderator, Employee, RapidMiner Certified Analyst, Community Manager, Member, University Professor, PM Moderator Posts: 2,959 Community Manager
    hi @tuan_nguyen I'd file a support ticket for this one. :wink:
    Scott
Sign In or Register to comment.