Generate a SQL implementation of your fitted linear model. This is currently targeted at TSQL, but may support other dialects in the future. This allows predictions to be made directly within the database without needing to query the data into an external R script. This package is redundant if you have R Services installed on SQL Server as you can call the R model directly.
Example
model <- lm(mpg ~ cyl + wt, data=mtcars)
summary(model)
Call:
lm(formula = mpg ~ cyl + wt, data = mtcars)
Residuals:
Min 1Q Median 3Q Max
-4.2893 -1.5512 -0.4684 1.5743 6.1004
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 39.6863 1.7150 23.141 < 2e-16 ***
cyl -1.5078 0.4147 -3.636 0.001064 **
wt -3.1910 0.7569 -4.216 0.000222 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 2.568 on 29 degrees of freedom
Multiple R-squared: 0.8302, Adjusted R-squared: 0.8185
F-statistic: 70.91 on 2 and 29 DF, p-value: 6.809e-12
Calling toSql(model, "dbo.PredictMpg") returns
CREATE FUNCTION dbo.PredictMpg (@cyl NUMERIC(24, 6), @wt NUMERIC(24, 6))
RETURNS NUMERIC(10,6)
BEGIN
-- Coefficient declarations
DECLARE @coefIntercept NUMERIC(24, 6) = 39.6862614802529;
DECLARE @coefCyl NUMERIC(24, 6) = -1.5077949682598;
DECLARE @coefWt NUMERIC(24, 6) = -3.19097213898374;
-- Response calculation
RETURN @coefIntercept + (@coefCyl * @cyl) + (@coefWt * @wt);
END
The function can then be used anywhere that you can use a UDF. Example
DECLARE @mtcars TABLE (
cyl INT,
wt DECIMAL(5,3)
);
INSERT INTO @mtcars VALUES (6, 2.62), (6, 2.875), (4, 2.32)
SELECT
cyl,
wt,
dbo.PredictMpg(cyl, wt) AS Prediction
FROM
@mtcars
This returns
(3 row(s) affected)
cyl wt Prediction
----------- --------------------------------------- ---------------------------------------
6 2.620 22.279144
6 2.875 21.465446
4 2.320 26.252026
(3 row(s) affected)
The results from this are identical to predict(model, newdata = mtcars)