svm-subset(1)

📦 version 3.24, in debian bookworm, debian trixie, ubuntu jammy, ubuntu noble, ubuntu resolute, ubuntu stonking

💖 svm-subset — a subset selection tool for LIBSVM 💖

okay bestie, listen. this is svm-subset, and it's giving "let me just take a lil bite of your dataset first" energy. it lives in the User Manuals section, it's a LIBSVM thing, and it is lowkey so useful. 🦋

✨ SYNOPSIS ✨

svm-subset [ -s method ] dataset number [ output1 ] [ output2 ]

like. that's the whole vibe. dataset in, number out, and optionally two little files to catch the pieces. 💅

🌸 DESCRIPTION 🌸

so here's the tea, bestie: training on big data is time consuming. like, so time consuming. sometimes you're not ready to commit to the whole thing and you should honestly just work on a smaller subset first. 🫶

the python script subset.py randomly selects a specified number of samples. and for classification data specifically, we serve you a stratified selection so the class distribution in your subset stays the same as the original. no cap, that's the whole point. ✨

🎀 OPTIONS 🎀

  • -s method — picks your selection method. two choices, pick your fighter:
    • 0 — stratified selection (classification only) (default). this is the one you get if you say nothing at all. main character default energy. 👑
    • 1 — random selection. chaos mode. fair enough. 🌈
  • output1 — the subset. if output1 is omitted, the subset will be printed on the screen instead. so like, it's not lost, it's just chillin' in your terminal. 🍓
  • output2 — the rest of data. everything that didn't make the cut goes here. it's not personal, it's just math. 💜

📁 FILES 📁

See svm-train(1) for the format of dataset. yes bestie, go read that one too, iykyk. 🧚

⭐ EXAMPLES ⭐

svm-subset heart_scale 100 file1 file2

from heart_scale 100 samples are randomly selected and stored in file1. all remaining instances are stored in file2. and that's it, that's the whole trick. ate. 🔥

💀 BUGS 💀

Please report bugs to the Debian BTS. if something is acting up, tell the Debian folks, not me, i'm just the messenger bestie. 😭

💖 AUTHOR 💖

Chih-Chung Chang, Chih-Jen Lin <cjlin@csie.ntu.edu.tw>, Chen-Tse Tsai <ctse.tsai@gmail.com> (packaging)

👑 SEE ALSO 👑

svm-train(1), svm-predict(1)

ngl, go read the SEE ALSO ones too. they're the squad. 💅

and that's a wrap, bestie! this page is from the Linux manual, dated DEC 2009. yeah, 2009, it's vintage, it's iconic, it's still serving. 💜✨

📦 where this page comes from