evolreg | R Documentation |
Identify the best linear, logistic or mixed regression model using an evolutionary approach.
evolreg(
data,
Y,
X = c(),
alpha = 0.05,
nvar = 0,
iter = 4000,
multix = TRUE,
interaction = TRUE,
multidiv = TRUE,
nbind = c(),
wash = TRUE,
NAfreq = 1,
family = "lm",
plot = FALSE,
verbose = TRUE,
fast = TRUE
)
data |
a dataframe. |
Y |
the y to predict. |
X |
vector of variables whose presence we want to force in the model. |
alpha |
0 to 1. If there are too many variables and the argument wash=TRUE, use this p-value threshold to eliminate the variables whose effect is too insignificant (Risk of eliminating the variables that will have an effect once transformed or in interaction). |
nvar |
Maximum number of variables in the model. A default value is proposed according to the number of individuals. |
iter |
Number of iterations. |
multix |
FALSE or TRUE. To allow variable variants (log, exp, polynomial, ^2). |
interaction |
FALSE or TRUE. To allow interactions between variables. |
multidiv |
FALSE or TRUE. To allow the synthesis of variables combining the ratio of one variable divided by another. |
nbind |
Number of simulated individuals in the population of models to be crossed in an evolutionary approach. A default value is proposed according to the number of variables. |
wash |
TRUE or FALSE.To select the best variables when there are too many. |
NAfreq |
from 0 to 1. NA part allowed in the variables. 1 by default (100% of NA tolerate). |
family |
"lm", "logical" or "lmer". Type of regression |
plot |
To visualize the evolution of the R2 of the models obtained after each crossing. |
verbose |
To display a summary of the intermediate models. |
fast |
Paramètre qui stoppe les itérations lorsque le R carré des modèles n'évolue plus de façon significative. |
A strongest possible regression model chosen from the available variables.
data(mtcars)
evolreg(mtcars,"mpg",plot=TRUE)
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.