RcppDL自动编码器'；重建的数据具有相同的值

我正在使用RcppDL库做一些实验。训练后，我使用原始数据集重建数据。但是，所有数据都具有相同的值

我的数据（第一列是id）：

重建后我得到了什么：

           [,1]     [,2] [,3]      [,4]       [,5] [,6]      [,7]       [,8]      [,9] [,10]     [,11] [,12]
 [1,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
 [2,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
 [3,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
 [4,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
 [5,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
 [6,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
 [7,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
 [8,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
 [9,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
[10,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
[11,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
[12,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
[13,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
[14,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
[15,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
[16,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
[17,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
[18,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
[19,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
[20,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
[21,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
[22,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
[23,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
[24,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
[25,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
[26,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
[27,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
[28,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
[29,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
[30,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
[31,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
[32,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
[33,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
[34,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
[35,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
[36,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
[37,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
[38,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
[39,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1
[40,] 0.7905348 0.999999    1 0.4721214 0.01729769    1 0.8770443 0.05453092 0.0353921     1 0.9994972     1

我的代码非常简单：

  da_obj <- Rda(x.new)
  setCorruptionLevel(da_obj, 0.01)
  setHiddenRepresentation(da_obj, 8)
  setTrainingEpochs(da_obj, 500)
  setLearningRate(da_obj, 0.002)
  train(da_obj)
  coord <- reconstruct(da_obj, x.new)

---

da_obj嗨，我现在也在看这个包，我得到了同样的结果。几乎相同的重建值。相关矩阵仅包含接近1…的值。。。。奇怪。该软件包的作者说该软件包已被弃用，建议使用mxnet。嗨，我现在也在看这个软件包，我得到了相同的结果。几乎相同的重建值。相关矩阵仅包含接近1…的值。。。。奇怪。该软件包的作者说该软件包已被弃用，并建议使用mxnet。
  da_obj <- Rda(x.new)
  setCorruptionLevel(da_obj, 0.01)
  setHiddenRepresentation(da_obj, 8)
  setTrainingEpochs(da_obj, 500)
  setLearningRate(da_obj, 0.002)
  train(da_obj)
  coord <- reconstruct(da_obj, x.new)