Vertica In-database analytics using R machine learning algorithm
Hello ,
I have a readily available R Prediction script with me that predicts with the help of Machine Learning algorithm written by me.
Now, I want to test its applicability in Vertica.
I have installed R-Vertica Lang package in my linux machine and also made everything ready to run.
My Question:
Is it possible to run the R-script in vertica for in-database predictive analytics.
Or
We should use those that are readily available only for use.
Note: I have tried many approaches but i am yet to get through it.
Also, I have successfully pulled data from Vertica into R and did prediction.
Now, I want to go reverse and run the R-script in vertica database.
Any help on this would be greatly appreciated.
Thank You.
Kalpesh
0
Comments
Hello Kalpesh,
Yes, you can write a prediction function as R UDX and run prediction in Vertica.
We have started to add machine learning aglorithms to Vertica from version 7.2.2. In 7.2.2, logistics regression, linear regression and kmeans are added. Their training functions, evaluation functions and prediction functions are all included. If you happen to use these algorithms, you don't need to develop them by yourself any more. These algorithms take advantage of Vertics MPP platform and run in a parallel and scalable fashion.
If you really like to use R, please try vertica.dplyr (https://github.com/vertica/vertica.dplyr), which allows you to work with Vertica from R.
Vincent
I work for HPE