![[Experimental]](figures/lifecycle-experimental.svg)
Exploratory analyses for linear regression models with multiple response, predictor, and two-way interaction variables. (lm models).
At the moment, multi-categorical variables are not supported as predictors or interactions (but control is fine). Binary variable should be numeric instead of factor
glm_model_explore(
data,
response_variable,
predictor_variable,
family,
two_way_interaction_variable = NULL,
three_way_interaction_variable = NULL,
control_variable = NULL,
marginal_alpha = 0.1,
verbose = TRUE,
show_p = TRUE,
return_result = FALSE,
print_control = FALSE,
plot_interaction = FALSE,
file_dir = NULL,
device = "jpeg",
width = 8.5,
height = 5,
units = "in",
y_lim = c(0, 1)
)Arguments
- data
data.frame- response_variable
Response variable. Support
dplyr::select()syntax.- predictor_variable
Pred variable. Support
dplyr::select()syntax.- family
a GLM family. It will passed to the family argument in glm See
?glmfor possible options.- two_way_interaction_variable
Two-way interaction variable. Each two-way interaction variable will interact with each pred variable. Support
dplyr::select()syntax.- three_way_interaction_variable
Three-way interaction variable. Each three-way interaction variable will interact with each pred and two-way interaction variables. Support
dplyr::select()syntax.- control_variable
Control variables. Support
dplyr::select()syntax.- marginal_alpha
Set marginal_alpha level for marginally significant (denoted by
.). Set to 0.05 if do not want marginally significant denotation.- verbose
Default is
TRUE. Set toFALSEto suppress outputs- show_p
Default is
TRUE. WhenTRUE, show the p-value in parenthesis.- return_result
Default is
FALSE. IfTRUE, it returns the model estimates as a data frame.- print_control
Default is
FALSE. IfTRUE, print coefficients of control variables.- plot_interaction
Default is
FALSE. IfTRUE, interactions will be plotted and saved on your laptop.- file_dir
Path of the directory to save plot to
- device
Default is "jpeg". See ggplot2::ggsave() for all options.
- width
Default is 8.5 (i.e., letter size width).
- height
Default is 5.
- units
Default is inches. Options are "in", "cm", "mm" or "px".
- y_lim
the plot's upper and lower limit for the y-axis. Length of 2. Example:
c(lower_limit, upper_limit)
Value
data.frame
Examples
test = data.frame(y1 = sample(c(0, 1), size = 1000, replace = TRUE),
y2 = sample(c(0, 1), size = 1000, replace = TRUE),
y3 = sample(c(0, 1), size = 1000, replace = TRUE),
x1 = rnorm(1000,100,10),
x2 = rnorm(1000,10,1),
x3 = rnorm(1000,6,2),
m1 = rnorm(1000,3,1),
m2 = rnorm(1000,2,0.5),
m3 = rnorm(1000,9,0.1),
c1 = rnorm(1000,5,0.4),
c2 = rnorm(1000,2,0.2),
c3 = rnorm(1000,7,0.9)
)
glm_model_explore(data = test,
response_variable = c(y1,y2,y3),
predictor_variable = c(x1,x2,x3),
two_way_interaction_variable = c(m1,m2,m3),
family = binomial(link = 'logit'),
control_variable = c(c1,c2,c3))
#> ──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────
#> Response Pred Interact_pred Interact_term Interact_term_coef Interact_pred_coef Pred_coef (Intercept) df r2
#> ──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────
#> y1 x1 m1 x1*m1 0.007 (0.245) -0.739 (0.257) -0.022 (0.286) 2.938 (0.206) 993.000 0.004
#> y1 x1 m2 x1*m2 -0.018 (0.190) 1.862 (0.179) 0.036 (0.194) -2.917 (0.331) 993.000 0.004
#> y1 x1 m3 x1*m3 -0.004 (0.946) -0.277 (0.966) 0.040 (0.945) 3.232 (0.955) 993.000 0.004
#> y1 x2 m1 x2*m1 -0.103 (0.112) 1.041 (0.109) 0.355 (0.079) + -2.889 (0.221) 993.000 0.006
#> y1 x2 m2 x2*m2 0.242 (0.063) + -2.344 (0.070) + -0.417 (0.107) 4.866 (0.081) + 993.000 0.007
#> y1 x2 m3 x2*m3 1.335 (0.042) * -14.121 (0.033) * -11.964 (0.043) * 127.413 (0.033) * 993.000 0.009
#> y1 x3 m1 x3*m1 0.002 (0.959) 0.003 (0.988) 0.023 (0.827) 0.622 (0.628) 993.000 0.003
#> y1 x3 m2 x3*m2 -0.044 (0.521) 0.314 (0.468) 0.113 (0.412) 0.030 (0.982) 993.000 0.004
#> y1 x3 m3 x3*m3 0.435 (0.163) -3.334 (0.094) + -3.884 (0.166) 30.603 (0.088) + 993.000 0.006
#> y2 x1 m1 x1*m1 0.000 (0.967) 0.042 (0.948) -0.001 (0.962) 0.020 (0.993) 993.000 0.002
#> y2 x1 m2 x1*m2 -0.016 (0.250) 1.852 (0.184) 0.031 (0.269) -3.398 (0.259) 993.000 0.006
#> y2 x1 m3 x1*m3 0.029 (0.652) -2.840 (0.658) -0.260 (0.651) 25.798 (0.655) 993.000 0.001
#> y2 x2 m1 x2*m1 0.201 (0.003) ** -1.938 (0.004) ** -0.651 (0.002) ** 6.634 (0.007) ** 993.000 0.012
#> y2 x2 m2 x2*m2 0.009 (0.947) 0.179 (0.889) -0.077 (0.764) 0.422 (0.879) 993.000 0.006
#> y2 x2 m3 x2*m3 -0.790 (0.223) 7.978 (0.223) 7.062 (0.227) -71.113 (0.227) 993.000 0.003
#> y2 x3 m1 x3*m1 -0.029 (0.375) 0.246 (0.243) 0.112 (0.277) -0.786 (0.541) 993.000 0.004
#> y2 x3 m2 x3*m2 0.021 (0.757) 0.126 (0.772) -0.019 (0.890) -0.182 (0.896) 993.000 0.005
#> y2 x3 m3 x3*m3 -0.285 (0.355) 1.766 (0.369) 2.590 (0.350) -15.841 (0.371) 993.000 0.002
#> y3 x1 m1 x1*m1 0.001 (0.904) -0.069 (0.915) 0.009 (0.659) -1.915 (0.409) 993.000 0.006
#> y3 x1 m2 x1*m2 0.008 (0.585) -0.558 (0.687) -0.003 (0.902) -1.012 (0.736) 993.000 0.008
#> y3 x1 m3 x1*m3 -0.095 (0.143) 8.947 (0.169) 0.866 (0.138) -82.659 (0.158) 993.000 0.008
#> y3 x2 m1 x2*m1 0.036 (0.572) -0.357 (0.576) -0.106 (0.593) 0.086 (0.971) 993.000 0.003
#> y3 x2 m2 x2*m2 -0.010 (0.937) 0.292 (0.819) 0.016 (0.950) -1.530 (0.579) 993.000 0.005
#> y3 x2 m3 x2*m3 -0.998 (0.125) 9.500 (0.147) 8.990 (0.125) -86.554 (0.143) 993.000 0.006
#> y3 x3 m1 x3*m1 0.028 (0.389) -0.170 (0.416) -0.059 (0.566) -0.633 (0.622) 993.000 0.004
#> y3 x3 m2 x3*m2 0.073 (0.284) -0.255 (0.556) -0.121 (0.381) -0.634 (0.648) 993.000 0.006
#> y3 x3 m3 x3*m3 0.375 (0.226) -2.767 (0.162) -3.351 (0.229) 23.697 (0.183) 993.000 0.005
#> ──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────
#> You can drag and resize the R console to view the entire table
#> Note: Coefficient (p-value): + p < 0.1, * p < 0.05, ** p < 0.01, *** p < 0.001