LIBSVM

What it does

LIBSVM accepts labeled training and test data in a sparse, indexed feature format, and also supports precomputed kernel values. Its training tools build support-vector models for multiclass classification, regression, or one-class estimation; users can then save a model and apply it to test examples. The package includes scaling and prediction tools, cross-validation, probability estimates, multiple kernel choices, and an automatic parameter-selection script. Outputs include model files, predicted class labels or regression values, and optional probability estimates.

Developers can use the library in their own software or run its training, scaling and prediction commands. The maintained distribution includes C++ and Java source and a Python interface; the project also links community interfaces whose maintenance and compatibility vary. Download the official release or install the Python package, prepare labeled sparse-format data and select model parameters. The optional easy.py workflow needs Python and gnuplot. Source and binary redistribution are permitted under the project’s copyright conditions; retain the required notices. The library has no hosted inference entitlement, so users supply their own execution environment.

What it can do

• Support-vector classification, including multiclass classification
• Support-vector regression
• One-class SVM distribution estimation
• Cross-validation for model selection
• Probability estimates
• Input scaling and automatic parameter selection

Works with

• Python interface
• R interface (e1071)
• MATLAB interface
• Java interface
• Perl interface
• Ruby interface
• Weka interface
• C# .NET interface

Before you choose

• Prepare and scale data, choose an appropriate kernel and validate parameters for the task.
• The optional easy.py workflow requires Python and gnuplot; community language bindings have their own compatibility limits.
• Retain the license notices and provide your own compute environment.

Listed under

• Coding
• Classification
• Developer library
• Machine learning
• Open source
• Regression
• Research

Get started

1. Download the LIBSVM archive or install the Python interface using `pip install -U libsvm-official`.
2. Read the included README and format labeled examples as indexed feature-value data.
3. Scale training and test features consistently, then train a model and use the prediction tool; for the automatic quick-start script, install Python and gnuplot.

Free software under the project copyright conditions; execution resources and integration work are separate.

LIBSVM is free to use. See their pricing page ↗

Sources for this overview

Product pages, documentation, and pricing checked 1 Nov 2026. Each source supports part of the overview; availability and prices can change.

• LIBSVM -- A Library for Support Vector Machines ↗www.csie.ntu.edu.tw
• ↗raw.githubusercontent.com
• GitHub - cjlin1/libsvm: LIBSVM -- A Library for Support Vector Machines · GitHub ↗github.com
• GitHub - ccerhan/LibSVMsharp: C# wrapper of LibSVM · GitHub ↗github.com

Company website

csie.ntu.edu.tw: product, pricing and policy pages

YC directory mirror

Batch and company name

Human review

Checked by a person

knowledge_base

manual_review

Last checked 1 Nov 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.