LIBLINEAR-PREDICT(1)

💖 LIBLINEAR-PREDICT 💖

okay bestie, welcome to the LIBLINEAR-PREDICT(1) page, living in the General Commands Manual section, and it's literally the moment your trained model gets to ✨shine✨. It's called liblinear-predict, and the command you actually type in the synopsis is linear-predict (yes, two names, it's giving mysterious 💅, just roll with it).

---

🌸 NAME 🌸

liblinear-predict — make predictions based on a trained linear classifier model 🧚

---

🎀 SYNOPSIS 🎀

linear-predict [options] test_file model_file output_file

---

🦋 DESCRIPTION 🦋

So here's the vibe: liblinear-predict takes your model-file (the linear classifier you already trained, no cap) and goes through every single sample in test_file, making a prediction for each one. Then it drops all those results into output_file. That's it. That's the whole job. And it ate. 🔥

---

✨ OPTIONS ✨

A lil summary of the options is right here, bestie:

• -b (0|1) — do you want probability estimates or not? <pink>0</pink> = nope, <pink>1</pink> = yes please. The default is <pink>0</pink>. 💖
• -q — quiet mode, aka no outputs, aka the "not me, I'm being mysterious" setting. 🤫

---

🌈 EXAMPLES 🌈

Wanna train a linear SVM using the L2-loss function? Use linear-train(1), like this:

liblinear-train data_file

And now for the main character moment — outputting probability estimates (for logistic regression only, iykyk):

liblinear-predict -b 1 test_file data_file.model output_file

bestie, note the <pink>1</pink> after -b — that's the "yes, probabilities please" signal. And <yellow>logistic regression only</yellow> is not a suggestion, it's the rules. 🙈

---

👑 SEE ALSO 👑

• liblinear-train(1)
• svm-predict(1)
• svm-train(1)
• svm-scale(1)

---

💜 AUTHORS 💜

liblinear-predict was written by the LIBLINEAR authors at National Taiwan university for the LIBLINEAR Project. Slay, honestly. 🫶

This manual page was written by Christian Kastner <ckk@debian.org>, for the Debian project (and may be used by others).

---

and that's the whole page, bestie! this version is from Ocotber 21, 2019 (yes, "Ocotber", we're keeping it, it's iconic 💀) and it's the LIBLINEAR-PREDICT(1) page — go predict something gorgeous. bye bestie, mwah 💋💖