考虑特定序列，将一组函数应用于多个数据帧

Question

考虑以下数据：

df <- structure(list(V1 = c(1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 
0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 
1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0), V2 = c(0, 0, 0, 0, 0, 0, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 0, 0), V3 = c(1, 
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 1, 1, 0, 1, 1, 
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 
1, 1, 1), V4 = c(1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
0, 0, 0, 0, 1, 1, 1, 0, 0, 0), V5 = c(1, 1, 0, 0, 0, 0, 0, 0, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1), V6 = c(0, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 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, 0, 0), V7 = c(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
0, 0, 1, 1, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 
1, 1, 1, 1, 0, 0, 0, 0, 0, 0), V8 = c(1, 1, 1, 0, 0, 0, 0, 1, 
1, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 0, 0), V9 = c(1, 
1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 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)), class = "data.frame", row.names = c(NA, -46L))

我想对df中的所有可能对应用以下函数。

应用于对V1和V2 ：

df$V1V2 <- (df$V1 * df$V2) + (1 - df$V1) * (1 - df$V2)
sum <- (sum(df$V1V2)/46)
df$VD <- (df$V1/sd(df$V1))
df$VI <- (df$V2/sd(df$V2))
est <- lm(df$VD ~ df$VI)
summary(est)
ndf <- data.frame(NA)
ndf$V1V2 <- summary(est)[["coefficients"]][, "Pr(>|t|)"][2]

应用于V1和V3对：

df$V1V3 <- (df$V1 * df$V3) + (1 - df$V1) * (1 - df$V3)
sum <- (sum(df$V1V3)/46)
df$VD <- (df$V1/sd(df$V1))
df$VI <- (df$V3/sd(df$V3))
est <- lm(df$VD ~ df$VI)
summary(est)
ndf$V1V3 <- summary(est)[["coefficients"]][, "Pr(>|t|)"][2]

我可以将此应用于所有其他对V1 （即V3 、 V4 、 V5 、 V6 、 V7 、 V8和V9 ）。 尽管如此，我确信这不是最好的方法。

更新

这是我走了多远：

dfV1 <- df[, c("V1", "V2", "V1", "V3", "V1", "V4", "V1", "V5", "V1", "V6", "V1", "V7", "V1", "V8", "V1", "V9")]
colnames(dfV1) <- c("V1", "V2", "V1", "V3", "V1", "V4", "V1", "V5", "V1", "V6", "V1", "V7", "V1", "V8", "V1", "V9")

sep <- lapply(seq(1, ncol(dfV1), by=2), function(i) 
  dfV1[i: pmin((i+1), ncol(dfV1))])

V1V2 <- sep[[1]]
V1V3 <- sep[[2]]
V1V4 <- sep[[3]]
V1V5 <- sep[[4]]
V1V6 <- sep[[5]]
V1V7 <- sep[[6]]
V1V8 <- sep[[7]]
V1V9 <- sep[[8]]

list_V1 <- tibble::lst(V1V2, V1V3, V1V4, V1V5, V1V6, V1V7, V1V8, V1V9)

library(dplyr)
my_func <- function(x) {
  x <- x %>%
    mutate(First = (x[,1] * x[,2] + (1 - x[,1] * (1 - x[,2]))),
           VD = x[,1] / sd(x[,1]),
           VI = x[,2] / sd(x[,2]))
}

res <- lapply(list_V1, my_func)
list2env(res, .GlobalEnv)

df.IC.V1 <- cbind.data.frame(V1V2$First, V1V3$First, V1V4$First, V1V5$First, V1V6$First, V1V7$First, V1V8$First, V1V9$First)
IC.all.V1 <- data.frame(colSums(df.IC.V1)/46)

我不知道如何将此部分应用于列表dfV1 ：

est <- lm(df$VD ~ df$VI)
summary(est)
ndf$V1V3 <- summary(est)[["coefficients"]][, "Pr(>|t|)"][2]

Answer 1

避免使用宽数据并保持数据长或整洁，这几乎有助于数据科学的各个方面：聚合、建模和绘图。 因此，考虑reshape （或pivot_longer的整洁语义）以将数据重新格式化为 long 并使用通用方法， by 9 个不同变量中的每一个调用来运行 model。

最后，您可能希望在新数据框中分离出此过程的结果，以避免由于长度不同而导致原始值重复，特别是因为您的 model 输出两行用于截距和多列变量。

long_df <- reshape(df, idvar="V1", varying=names(df)[-1], 
                   times=names(df)[-1],
                   v.names="value", timevar="variable", 
                   new.row.names=1:1E4, direction="long")

pairwise_model <- function(sub) {
   sub$vpair <- (sub$V1 * sub$value) + (1 - sub$value) * (1 - sub$value)
   sum_v <- (sum(sub$vpair)/46)
   sub$VD <- (sub$V1/sd(sub$V1))
   sub$VI <- (sub$value/sd(sub$value))
   est <- lm(VD ~ VI, data=sub)
   print(summary(est))

   ndf <- data.frame(variable = sub$variable[[1]],
                     summary(est)[["coefficients"]])
   return(ndf)
}

df_list <- by(long_df, long_df$variable, pairwise_model)

results_df <- do.call(rbind, df_list) 

results_df
#                variable    Estimate Std..Error    t.value     Pr...t..
# V2.(Intercept)       V2  0.70897157  0.1697361  4.1769050 1.376794e-04
# V2.VI                V2  0.20546559  0.1475392  1.3926169 1.707334e-01
# V3.(Intercept)       V3  0.00000000  0.4324256  0.0000000 1.000000e+00
# V3.VI                V3  0.29294628  0.1441419  2.0323471 4.818178e-02
# V4.(Intercept)       V4  0.59719677  0.1461565  4.0860091 1.829775e-04
# V4.VI                V4  0.47663445  0.1325296  3.5964390 8.118808e-04
# V5.(Intercept)       V5  0.77259722  0.1604268  4.8158874 1.766754e-05
# V5.VI                V5  0.13627939  0.1493492  0.9124883 3.664845e-01
# V6.(Intercept)       V6  0.68720490  0.1557572  4.4120255 6.533736e-05
# V6.VI                V6  0.31399751  0.1431310  2.1937768 3.357833e-02
# V7.(Intercept)       V7  0.57392936  0.1516681  3.7841150 4.627413e-04
# V7.VI                V7  0.46128020  0.1337587  3.4486009 1.253140e-03
# V8.(Intercept)       V8  0.90047538  0.1869717  4.8161045 1.765498e-05
# V8.VI                V8 -0.09345783  0.1500959 -0.6226544 5.367256e-01
# V9.(Intercept)       V9  0.60871296  0.1645124  3.7001029 5.960399e-04
# V9.VI                V9  0.35598290  0.1408800  2.5268512 1.517916e-02

在线演示