Note that there are some explanatory texts on larger screens.

plurals
  1. POHow do I obtain the eigenvalues of a huge matrix (size: 2x10^5)
    primarykey
    data
    text
    singulars
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    plurals
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. COIn double precision just storing the dense 200000x200000 matrix will be ~300GB of memory. The problem is most likely not the programming language that you use, but the RAM and swap space of the computer you are using.
      singulars
    2. CODo you know the nature of your matrix? If it's a random Gaussian matrix (or if you can approximate your data to be so), there are several analytical results I could provide that will tell you the distribution of the eigenvalues (asymptotically). For a matrix of this size, you almost _cannot_ go wrong if you use the asymptotic results, and I'd dare to wager that a 200000 point sample from that distribution will be pretty damn near the actual eigenvalues. If this is not the case, do you know if your requirements are small? Like for e.g., do you need just the largest eigenvalue or all?
      singulars
      1. This table or related slice is empty.
    3. COFurther along yoda's line of thought: if you're doing stats work (which includes a lot of what is called data mining and machine learning), do you really need to work with the full data set or would a smaller (~1e4) sample be sufficient? You might consider providing some details about the problem that is behind this matrix, there might be better ways of getting what you're after.
      singulars
 

Querying!

 
Guidance

SQuiL has stopped working due to an internal error.

If you are curious you may find further information in the browser console, which is accessible through the devtools (F12).

Reload