I apologize ahead of time if I have missed something simple, but for the life of me, I don’t see an easy explanation of how to deploy a model. Loading data (no problem), transforming data( piece of cake), testing models and performance (got it)……. OK now I want to USE IT IN THE REAL WORLD (well duh, I have no clue based on what I have seen, there seems to be a jump or gap in-between that is not being explained)
Heres what I did.
I have created a linear regression (yes simple) model that predicts next Fridays S&P 500 closing price index based on this week’s data. Cool, fun Backtesting.
I want to use this in the real world now, not for testing anymore.
OK, I have this weeks data (its Friday) and want to plug it in for my next Friday prediction (based on the model I created and tested). What the heck do I do?
I don’t need it deployed to the web for others, I just want it here, for me on my local machine so that I can place trades based on it. What do I do? All I saw was a big jump to unknown terms about PMML predictor and dedicated predictor with the distinct impression that somebody accidentally skipped from chapter 3 to chapter 6 and assumed that we knew chapters 4 and 5!
Can somebody please give a “for dummies” version of how to deploy a simple linear regression into real-time production and generate my desired future outputs (or point to a good SIMPLE explanation of one?) Please don’t suggest this one https://www.knime.com/blog/seven_modes_deployment . I’ve been there and found it completely worthless and was more confused after watching it than before.
This is occasionally where free software falls down, I would have gladly paid to pick up the phone and call support. I suppose I should check, maybe KNIME has paid support plans. The answer is probably out there if I look long and hard enough, it’s just that I’m involved in 15 (exaggerating) other projects and I can’t take a whole day or two trying to find the answer to this seemingly simple question. Frustrated…arghhhhh
Thank you so much! (for any help and or letting me vent a bit)