Question

我有一个函数，我从sum函数获得NA，第一个sum效果很好，但第二个sum不起作用并返回NA。

这是功能：

  Gramm.Pred.Err <- function(acts , grammProbs)
  {
    acts <- as.numeric(acts)
    grammProbs <-  as.numeric(grammProbs)

    print("acts is:")
    print(acts)
    print("grammProbs is:")
    print(grammProbs)


    false.ind = grammProbs == 0
    hit.ind= grammProbs>0

    print("false.ind is:")
    print(false.ind)
    print("hit.ind is:")
    print(hit.ind)


    hit.acts = acts[hit.ind]
    hit.probs = grammProbs[hit.ind]

    print("hit.acts is:")
    print(hit.acts)
    print("hit.probs is:")
    print(hit.probs)

    misses.ind = hit.acts< hit.probs

    hits =  sum(hit.acts)

    falses =  sum( acts[false.ind])

    misses = sum( hit.probs[misses.ind] - hit.acts[misses.ind] )

    print("misses.ind is:")
    print(misses.ind)
    print("hits is:")
    print(hits)
    print("falses is:")
    print(falses)
    print("misses is:")
    print(misses)

    print("final:")
    print(1-(hits/(hits+falses+misses)))

    return (1-(hits/(hits+falses+misses)))


  }

act和grammProbs矢量：

activations.vector <- c(  "2.08101344",  "-1.41434467",   "0.07817803",   "0.45509970",   "1.27916718",  "-0.84691423",   "2.01260424",  "-1.42960405",  "-1.47239423",
   "0.68798345",  "-0.86126810",   "1.61871290",  "-1.49541676",   "3.70249152",   "1.60749793",  "-2.68202949",   "0.58367389",  "-0.15213574",
  "9.20287609",   "1.62072563",   "4.33229876",  "-0.16497207",  "-0.16517217", "-13.28754520")

prob.vector <- c( "2.0000000", "0.0000000", "0.1666670", "0.1666670", "0.0833333", "0.0833333", "0.0833333", "0.0833333", "0.0000000", "0.0000000", "0.0833333", "0.0833333",
 "0.0000000", "0.0000000", "0.0833333", "0.0833333", "0.0000000", "0.0000000", "0.0000000", "0.0000000", "0.0000000", "0.0000000", "0.0000000", "0.0000000")

对于这些，我得到以下结果：

gpe.matrix <- matrix(, 10, 24)

gpe.matrix[1,] <- Gramm.Pred.Err(acts = activations.vector,grammProbs = prob.vector)

[1] "acts is:"
 [1]   2.08101344  -1.41434467   0.07817803   0.45509970   1.27916718  -0.84691423   2.01260424  -1.42960405  -1.47239423
[10]   0.68798345  -0.86126810   1.61871290  -1.49541676   3.70249152   1.60749793  -2.68202949   0.58367389  -0.15213574
[19]   9.20287609   1.62072563   4.33229876  -0.16497207  -0.16517217 -13.28754520
[1] "grammProbs is:"
 [1] 2.0000000 0.0000000 0.1666670 0.1666670 0.0833333 0.0833333 0.0833333 0.0833333 0.0000000 0.0000000 0.0833333 0.0833333
[13] 0.0000000 0.0000000 0.0833333 0.0833333 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000
[25] 0.0000000
[1] "false.ind is:"
 [1] FALSE  TRUE FALSE FALSE FALSE FALSE FALSE FALSE  TRUE  TRUE FALSE FALSE  TRUE  TRUE FALSE FALSE  TRUE  TRUE  TRUE  TRUE
[21]  TRUE  TRUE  TRUE  TRUE  TRUE
[1] "hit.ind is:"
 [1]  TRUE FALSE  TRUE  TRUE  TRUE  TRUE  TRUE  TRUE FALSE FALSE  TRUE  TRUE FALSE FALSE  TRUE  TRUE FALSE FALSE FALSE FALSE
[21] FALSE FALSE FALSE FALSE FALSE
[1] "hit.acts is:"
 [1]  2.08101344  0.07817803  0.45509970  1.27916718 -0.84691423  2.01260424 -1.42960405 -0.86126810  1.61871290  1.60749793
[11] -2.68202949
[1] "hit.probs is:"
 [1] 2.0000000 0.1666670 0.1666670 0.0833333 0.0833333 0.0833333 0.0833333 0.0833333 0.0833333 0.0833333 0.0833333
[1] "misses.ind is:"
 [1] FALSE  TRUE FALSE FALSE  TRUE FALSE  TRUE  TRUE FALSE FALSE  TRUE
[1] "hits is:"
[1] 3.312458
[1] "falses is:"
[1] NA
[1] "misses is:"
[1] 6.241638
[1] "final:"
[1] NA

非常感谢您的建议。

Answer 1

我逐行查看了你的功能，发现了问题：

falses = sum（act [false.ind]）

acts是长度为24的向量，false.ind是长度为25的向量。因此，您正在尝试对不存在的向量元素进行子集化。这会产生NA。

如果你想得到没有NA的向量之和，你可以在总结这样的向量元素之前删除NA：

falses = sum（act [false.ind]，na.rm = T）

来自R中的数字之和的NA输出

1 个答案: