0

I am a beginner to statistics and was wondering if it is possible to estimate the density where there are 50 input variables.

Do I perform density estimation for each variable individually then multiply them or do I consider the density of a vector of length 50 of the variables?

Gabi23
  • 143
  • 1
    What comes after density estimation? What is the reason you're doing it? – BruceET Sep 07 '21 at 15:54
  • What exactly are you trying to do? What is your data? Could you give us an example, e.g. few rows of the data? – Tim Sep 07 '21 at 18:19
  • @BruceET My goal is to identify pieces of data that are anomalies and see if these anomalies are responsible for anomalies in the output predicted by the neural network. – Gabi23 Sep 07 '21 at 19:55
  • @Tim I read a research paper that using a list of features predicted distances between zeros of a function using a neural network. They suggested that someone could maybe estimate to density of the input data and see if anomalous input data could help identify anomalous distances between zeros. I thought to give this a try but I’ve no background in statistics, hence my question. – Gabi23 Sep 07 '21 at 19:58
  • @Tim Here is a few pieces of the data (there are nearly 100000 in total) (note: each component is separately rescaled between -1 and 1): – Gabi23 Sep 07 '21 at 20:06
  • @Tim {0.324683, -0.0609654, 0.00464909, 0.0332194, 0.0411025, -0.0673522, 0.0629532, 0.000554243, 0.0551618, -0.0674554, 0.0592887, 0.144113, 0.239823, -0.182317, 0.0555712, 0.0222715, -0.0634697, -0.0191774, -0.00308826, 0.0608748, -0.113239, 0.0849235, -0.41635, 0.565762, 0.903525, -0.359827, 0.899794, 0.864537, 0.183262, -0.625577, -0.522097, -0.081364, 0.162094, -0.264774, -0.0559528, 0.0286839, 0.151159, -0.0183345, 0.197464, -0.78723, -0.238502, -0.257547, -0.15483, -0.201061, 0.117934, -0.345785, -0.404379, -0.0750519, -0.0200682, -0.184341} – Gabi23 Sep 07 '21 at 20:07
  • @Tim {-0.326954, 0.049429, 0.0132058, 0.122221, 0.0219107, 0.087662, 0.0129931, 0.0864528, 0.0369861, -0.058427, 0.12637, -0.167936, -0.293847, 0.102893, 0.0196797, 0.0996321, 0.0275138, 0.037579, 0.0531149, 0.0178497, 0.0873258, -0.555409, -0.369951, -0.82301, -0.457824, -0.354692, -0.418983, -0.881794, -0.326863, -0.41396, -0.109965, -0.432755, -0.0627848, -0.224552, 0.0588479, -0.125872, 0.312467, 0.443384, -0.43074, -0.585926, -0.00685743, -0.168424, -0.221087, -0.134182, -0.187003, -0.0857966, -0.214191, -0.202227, 0.177754, -0.0772631} – Gabi23 Sep 07 '21 at 20:08
  • @Tim {0.069053, -0.0494288, 0.201197, -0.0905426, 0.27203, -0.00226303, 0.061733, -0.044544, 0.0379866, 0.0305012, 0.0800589, 0.064654, -0.0799105, 0.133991, 0.0150846, 0.0684807, 0.229635, -0.0621644, 0.0276965, -0.0664962, 0.0276059, 0.0475249, -0.134386, 0.314657, -0.888361, -0.801982, 0.656181, 0.156295, -0.588288, 0.0919509, -0.100751, -0.0361733, -0.432358, -0.372894, -0.142872, -0.485441, -0.25231, 0.251589, 0.386007, 0.15291, -0.270958, -0.0931162, 0.107068, -0.538917, -0.098557, -0.275693, -0.101589, -0.0564258, -0.03196, -0.521241} – Gabi23 Sep 07 '21 at 20:08
  • Please do not give new information only in comments, edit your question to add the new information. We want posts to be self-contained, comments can be deleted, and anyhow, information in comments are not well organized. Also, many people do not read comments. – kjetil b halvorsen Sep 08 '21 at 05:10

0 Answers0