extract.coef.xgb.Booster

Extract Coefficient Information from Models

# S3 method for xgb.Booster
extract.coef(model, feature_names = NULL,
  removeNonSelected = TRUE, ...)

Arguments

model

Model object from which to extract information.
feature_names

Names of coefficients
removeNonSelected

If TRUE (default) do not return the non-selected (0) coefficients
...

Further arguments

Value

A data.frame containing the coefficient, the standard error and the variable name.

Details

Gets the coefficient values and variable names from a model. Since xgboost does not have standard errors, those will just be NA.

Examples

library(xgboost)
data(diamonds, package='ggplot2')
diaX <- useful::build.x(price ~ carat + cut + x, data=diamonds, contrasts=FALSE)
diaY <- useful::build.y(price ~ carat + cut + x, data=diamonds)
xg1 <- xgboost(data=diaX, label=diaY,
booster='gblinear',
objective='reg:linear', eval_metric='rmse',
nrounds=50
)
#> [1]	train-rmse:3431.926270 
#> [2]	train-rmse:3110.120850 
#> [3]	train-rmse:2834.771729 
#> [4]	train-rmse:2602.429932 
#> [5]	train-rmse:2406.225098 
#> [6]	train-rmse:2241.269775 
#> [7]	train-rmse:2103.484863 
#> [8]	train-rmse:1989.205078 
#> [9]	train-rmse:1895.113037 
#> [10]	train-rmse:1818.168701 
#> [11]	train-rmse:1755.667480 
#> [12]	train-rmse:1705.199341 
#> [13]	train-rmse:1664.665649 
#> [14]	train-rmse:1632.252075 
#> [15]	train-rmse:1606.439453 
#> [16]	train-rmse:1585.940308 
#> [17]	train-rmse:1569.709351 
#> [18]	train-rmse:1556.883301 
#> [19]	train-rmse:1546.755493 
#> [20]	train-rmse:1538.773560 
#> [21]	train-rmse:1532.486816 
#> [22]	train-rmse:1527.534302 
#> [23]	train-rmse:1523.632690 
#> [24]	train-rmse:1520.557983 
#> [25]	train-rmse:1518.135010 
#> [26]	train-rmse:1516.224365 
#> [27]	train-rmse:1514.712769 
#> [28]	train-rmse:1513.516724 
#> [29]	train-rmse:1512.565674 
#> [30]	train-rmse:1511.812622 
#> [31]	train-rmse:1511.209839 
#> [32]	train-rmse:1510.725586 
#> [33]	train-rmse:1510.338745 
#> [34]	train-rmse:1510.024536 
#> [35]	train-rmse:1509.764771 
#> [36]	train-rmse:1509.555664 
#> [37]	train-rmse:1509.381226 
#> [38]	train-rmse:1509.228882 
#> [39]	train-rmse:1509.106812 
#> [40]	train-rmse:1508.997559 
#> [41]	train-rmse:1508.907593 
#> [42]	train-rmse:1508.827148 
#> [43]	train-rmse:1508.754150 
#> [44]	train-rmse:1508.691162 
#> [45]	train-rmse:1508.631348 
#> [46]	train-rmse:1508.575073 
#> [47]	train-rmse:1508.521484 
#> [48]	train-rmse:1508.474243 
#> [49]	train-rmse:1508.427124 
#> [50]	train-rmse:1508.381714 
extract.coef(xg1)
#> # A tibble: 8 x 3
#>     Value    SE Coefficient 
#>     <dbl> <dbl> <chr>       
#> 1  8011.     NA carat       
#> 2 -3168.     NA cutFair     
#> 3 -2050.     NA cutGood     
#> 4 -1724.     NA cutPremium  
#> 5 -1657.     NA cutVery Good
#> 6 -1373.     NA cutIdeal    
#> 7  -135.     NA (Intercept) 
#> 8   -72.7    NA x           
extract.coef(xg1, feature_names=colnames(diaX))
#> # A tibble: 8 x 3
#>     Value    SE Coefficient 
#>     <dbl> <dbl> <chr>       
#> 1  8011.     NA carat       
#> 2 -3168.     NA cutFair     
#> 3 -2050.     NA cutGood     
#> 4 -1724.     NA cutPremium  
#> 5 -1657.     NA cutVery Good
#> 6 -1373.     NA cutIdeal    
#> 7  -135.     NA (Intercept) 
#> 8   -72.7    NA x