Github user yb33 commented on the pull request: https://github.com/apache/spark/pull/1290#issuecomment-65342521 Hi guys, I exchanged a couple of emails offline with Alexander. Per his request I will start running some additional benchmark tests on the other data sets I got (i.e. advertising data, etc). My co-worker Girish will also join. Meanwhile, I have a question: how should the MLLib NN behave when it encounters some missing values in the input data (which is a very typical situation for industry/commercial data, including some of my data sets)? There are several possibilities, including but not limited to: 1. Quit as soon as any value is missing 2. Ignore the row that has any missing values, but continue to the next rows. 3. Replace missing values with user-specified (or hard-coded, such as zero, for example) defaults and continue. 4. More complex possibilities (to somehow use all data that is not missing, even partial row data). I suggest option (2) for now, what do you think?
--- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org