使用 R 创建子样本回归系数的森林图

Question

我有一个由 100 组组成的国际象棋位置数据集，每组占据 50 个位置之一（“Position_number”）和两种颜色之一（“stm_white”）。我想运行每个 Position_number 子样本的线性回归，其中 stm_white 是解释变量， stm_perform 是结果变量。然后，我想在森林图中显示 stm_white 的系数和每个回归的相关置信区间。这个想法是为了能够轻松地看到哪个 Position_number 子样本为 stm_white 提供了重要的系数，并比较不同位置的系数。例如，该图将有 50 个 y 轴类别，每个类别都标有每个位置编号，x 轴将表示系数范围，并且该图将为每个位置编号显示一个水平置信条。

我被困在哪里：

获取每个回归的置信区间界限
在一个图上绘制 50 个系数（带有置信区间）中的每一个。（我觉得这叫森林地块？）

这就是我目前获取每个回归系数列表的方式：

fits <- by(df, df[,"Position_number"],
           function(x) lm(stm_perform ~ stm_white, data = x))
# Combine coefficients from each model
do.call("rbind", lapply(fits, coef))

这里是 10 个职位的示例（如果有更好的方式来显示可重现的数据，我们深表歉意）：

>dput(droplevels(dfMWE[,c("Position_number","stm_white","stm_perform")]))
structure(list(Position_number = c(0, 0, 0, 0, 0, 0, 0, 0, 0, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 
2, 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 
3, 3, 3, 3, 3, 3, 3, 3, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 
4, 4, 4, 4, 4, 4, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 
5, 5, 5, 5, 5, 5, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 
6, 6, 6, 6, 6, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 
7, 7, 7, 7, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 
8, 8, 8, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 
9, 9, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 
10, 10, 10, 10, 10, 10), stm_white = c(0, 0, 0, 0, 0, 0, 0, 0, 
0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 
1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 
1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 
1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 
1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 
1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 
1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 
1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 
1, 1), stm_perform = c(0.224847134350316, -0.252000458803946, 
0.263005239459311, -0.337712202569111, 0.525880930891169, -0.5, 
0.514387184165999, 0.520136722035817, -0.471249436107731, -0.557311633762293, 
-0.382774969095054, -0.256365477992672, -0.592466230584332, 0.420100239642119, 
0.35728693116738, -0.239203909010858, 0.492804918290949, -0.377349804212738, 
0.498560888290847, 0.650604627933873, 0.244481117928803, 0.225852022298169, 
0.448376452689039, 0.305090287270497, 0.275461757157464, 0.0232950364735793, 
-0.117225030904946, 0.103523492101814, 0.098301745397805, 0.435599509759579, 
-0.323024628921732, -0.790798102797238, 0.326223812111678, -0.331305043692668, 
0.300230596737942, -0.340292005855252, 0.196181480575316, -0.0606495585093978, 
0.789844179758131, -0.0862623926308338, -0.560150145231903, 0.697345078589853, 
-0.425719796345476, 0.65321716721887, -0.878090073942596, 0.393712176214572, 
0.636076899687882, 0.530184680003902, -0.567228844342952, 0.767024918145021, 
-0.207303615824231, -0.332581578126777, -0.511510891217792, 0.227871326531416, 
-0.0140876421179904, -0.891010911045765, -0.617225030904946, 
-0.335142021445235, -0.517262524432376, 0.676301669492737, 0.375998241382333, 
-0.0882899718631629, -0.154706189382, -0.108431333126633, 0.204584592662721, 
0.475554538879339, 0.0840205872617279, -0.403370826694226, -0.74253555894307, 
0.182570385474772, -0.484175014735265, -0.332581578126777, -0.427127748605496, 
0.474119069108831, -0.0668284645696687, -0.0262098994728823, 
-0.255269593134965, -0.313699742316688, -0.485612815834001, 0.302654921410147, 
-0.425719796345476, 0.65321716721887, 0.393712176214572, 0.60766106412682, 
0.530184680003902, 0.384135895746244, 0.564400490240421, 0.767024918145021, 
0.702182602090521, 0.518699777929559, -0.281243170101218, -0.283576305897061, 
0.349395372066127, -0.596629173305774, 0.0849108889395813, -0.264122555898524, 
0.593855385236178, -0.418698521631085, 0.269754586702576, -0.719919005947152, 
0.510072446927438, -0.0728722513945044, -0.0849108889395813, 
0.0650557537775339, 0.063669188530584, -0.527315973006493, -0.716423694102939, 
-0.518699777929559, 0.349395372066127, -0.518699777929559, 0.420100239642119, 
-0.361262250888275, 0.431358608116332, 0.104596852632671, 0.198558626418023, 
0.753386077785615, 0.418698521631085, -0.492804918290949, -0.636076899687882, 
-0.294218640287997, 0.617225030904946, -0.333860575416878, -0.544494573083008, 
-0.738109032540419, -0.192575818328721, -0.442688366237707, 0.455505426916992, 
0.13344335621046, 0.116471711943561, 0.836830966002895, -0.125024693001636, 
0.400603203290743, -0.363923100312118, -0.157741327529574, -0.281243170101218, 
-0.326223812111678, -0.548774335859742, 0.104058949158278, -0.618584122089031, 
-0.148779202375097, -0.543066492022212, -0.790798102797238, -0.541637702714763, 
0.166337530816562, -0.431358608116332, -0.471249436107731, -0.531618297828107, 
-0.135452994588696, 0.444109038883147, -0.309993792719686, 0.472684026993507, 
-0.672509643334985, -0.455505426916992, -0.0304828450187082, 
-0.668694956307332, 0.213036720610531, -0.370611452782498, -0.100361684849949, 
-0.167940159469667, -0.256580594295053, 0.41031649686005, 0.544494573083008, 
-0.675040201040299, 0.683816314193659, 0.397841906825283, 0.384135895746244, 
0.634743335052317, 0.518699777929559, -0.598013765769344, -0.524445461120661, 
-0.613136820153143, 0.12949974225673, -0.337712202569111, -0.189904841395243, 
0.588289971863163, 0.434184796930767, -0.703385003471829, 0.505756208411145, 
0.445530625978324, -0.167137309739621, 0.437015271896404, -0.550199353253537, 
-0.489927553072562, -0.791748837508184, 0.434184796930767, 0.264122555898524, 
-0.282408276808469, -0.574280203654524, 0.167940159469667, -0.439849854768097, 
-0.604912902007957, 0.420100239642119, 0.35728693116738, 0.239220254140668, 
-0.276612130560829, -0.25746444105693, 0.593855385236178, -0.632070012100074, 
0.314483587504712, 0.650604627933873, -0.226860086923233, -0.702182602090521, 
0.25746444105693, -0.174474012638818, 0.0166045907672774, 0.535915926945102, 
0.141635395826102, 0.420100239642119, 0.557311633762293, 0.593855385236178, 
0.6961287704296, 0.0444945730830079, -0.234005329233511, 0.448376452689039, 
-0.86655664378954, 0.22107824319756, 0.148051654147426, 0.543066492022212, 
-0.448376452689039, 0.373300918333268)), row.names = c(NA, -220L
), groups = structure(list(Position_number = c(0, 1, 2, 3, 4, 
5, 6, 7, 8, 9, 10), .rows = structure(list(1:20, 21:40, 41:60, 
    61:80, 81:100, 101:120, 121:140, 141:160, 161:180, 181:200, 
    201:220), ptype = integer(0), class = c("vctrs_list_of", 
"vctrs_vctr", "list"))), row.names = c(NA, 11L), class = c("tbl_df", 
"tbl", "data.frame"), .drop = TRUE), class = c("grouped_df", 
"tbl_df", "tbl", "data.frame"))

Answer 1

confint() 可以得到模型的置信区间。
forestplot() 来自 forestplot R 包可以让你成为森林图。

library(dplyr)
library(forestplot)

results <- lapply(unique(df$Position_number), function(pos) {
    fit = filter(df, Position_number == pos) %>% 
        lm(data = ., stm_perform ~ stm_white)
    stm_white_lm_index = 2 # the second term in lm() output is "stm_white"
    coefficient = coef(fit)[stm_white_lm_index]
    lb = confint(fit)[stm_white_lm_index,1] # lower bound confidence 
    ub = confint(fit)[stm_white_lm_index,2] # upper bound confidence
    output = data.frame(Position_number = pos, coefficient, lb, ub)
    return(output)
}) %>% bind_rows() # bind_rows() combines output from each model in the list

with(results, forestplot(Position_number, coefficient, lb, ub))

森林图在左侧显示“Position_number”标签，并绘制了“stm_white”的回归系数和 95% 的置信区间。您可以进一步自定义绘图。有关详细信息，请参阅 Max Gordon 的 forestplot::forestplot() 或 this introduction。

使用 R 创建子样本回归系数的森林图

Using R to create forest plots of coefficients for regressions on subsamples

r

lm

forestplot