In tidymodels/modeldb: Fits Models Inside the Database

modeldb

library(dplyr)
library(modeldb)

Fit models inside the database! modeldb works with most database back-ends because it leverages dplyr and dbplyr for the final SQL translation of the algorithm. It currently supports:

K-means clustering
Linear regression

Installation

Install the CRAN version with:

install.packages("modeldb")

The development version is available from GitHub using remotes:

# install.packages("remotes")
remotes::install_github("tidymodels/modeldb")

Linear regression

An easy way to try out the package is by creating a temporary SQLite database, and loading mtcars to it.

con <- DBI::dbConnect(RSQLite::SQLite(), path = ":memory:")
RSQLite::initExtension(con)
dplyr::copy_to(con, mtcars)

library(dplyr)

tbl(con, "mtcars") %>%
  select(wt, mpg, qsec) %>%
  linear_regression_db(wt)

The model output can be parsed by tidypredict to run the predictions in the database. Please see the "Linear Regression" article to learn more about how to use linear_regression_db()

K Means clustering

To use the simple_kmeans_db() function, simply pipe the database back end table to the function. This returns a list object that contains two items:

A sql query table with the final center assignment
A local table with the information about the centers

km <- tbl(con, "mtcars") %>%
  simple_kmeans_db(mpg, wt)

colnames(km)

The SQL statement from tbl can be extracted using dbplyr's remote_query()

dbplyr::remote_query(km)

Contributing

This project is released with a Contributor Code of Conduct. By contributing to this project, you agree to abide by its terms.

For questions and discussions about tidymodels packages, modeling, and machine learning, please post on Posit Community.
If you think you have encountered a bug, please submit an issue.
Either way, learn how to create and share a reprex (a minimal, reproducible example), to clearly communicate about your code. Check out this helpful article on how to create reprexes for problems involving a database.
Check out further details on contributing guidelines for tidymodels packages and how to get help.