# cjlin1/liblinear

LIBLINEAR -- A Library for Large Linear Classification

Repository: https://github.com/cjlin1/liblinear
Canonical: https://ross.abutalabs.com/products/liblinear
Homepage: https://www.csie.ntu.edu.tw/~cjlin/liblinear/
Language: C++
License: BSD-3-Clause
License Family: permissive
Last push: 2025-12-29T11:32:11+00:00

## Health v2 (maintenance only)
Score: 59/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 59, release rhythm 35, longevity 100
- inputs: {"age_days": 4490, "days_push": 247, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1051, forks 342 (observed 2026-08-28T04:03:23.207967+00:00)

## What it is
LIBLINEAR is a C++ library and CLI toolset for training large-scale regularized linear classifiers, regressors, and one-class SVMs. It ships train/predict command-line programs plus Python and MATLAB/Octave interfaces, using the same sparse data format as LIBSVM.

## Use cases
- train a linear SVM on millions of text documents
- classify large sparse high-dimensional feature data
- fit logistic regression with L1 or L2 regularization
- run support vector regression on big datasets
- detect outliers with one-class SVM
- cross-validate and auto-select C parameters for linear models

## When to choose
- your dataset is large and linear models suffice, e.g. document classification
- you need fast training on millions of instances and features
- you want LIBSVM-compatible sparse data format with a simpler linear solver
- you need probability outputs via logistic regression

## When to avoid
- your data is small or you are a beginner - use LIBSVM first
- you need nonlinear kernel mappings for accuracy
- your data is unscaled or C is large, which can make the default solver slow

## Facets
- artifact type: library
- maturity: stable
- function: machine-learning
- domain: machine-learning, data-science
- platform: windows, python, cpp, cross-platform, cli
- tags: svm, logistic-regression, linear-classification, large-scale-learning, libsvm-format, matlab-interface, natural-language-processing, linux, macos

## Member repositories
- cjlin1/liblinear (main) score 59

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:23.207967+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T07:00:03.070080+00:00, confidence not recorded.
  - readme: https://github.com/cjlin1/liblinear (fetched 2026-08-28T04:03:23.207967+00:00, sha 3ffa03870fd9)
  - homepage: https://www.csie.ntu.edu.tw/~cjlin/liblinear/ (fetched 2026-08-29T13:01:20.265706+00:00, sha 64d8deefc72d)
  - site_page: https://www.csie.ntu.edu.tw/~cjlin/liblinear/FAQ.html (fetched 2026-08-29T13:01:20.268507+00:00, sha f336df15ec15)
- Data as of 2026-08-30T08:39:29.467469+00:00.