Welcome to orsvm’s documentation!

orsvm is a free software package which provides a SVM classifier with some novel orthogonal polynomial kernels. This library provides a complete path of using the SVM classifier from normalization to calculation of SVM equation and the final evaluation. Convext optimization is done using cvxopt, which solves the convext SVM’s equation and returns the support vectors as the result. orsvm benefits from the orthogonal kernels’ capabilities to constitute the kernel matrix, which is thereafter used by cvxopt to as one of input matrices , to form the SVM equation. Another novelty in orsvm is that it is now possible to transform the dataset to a fractional space, as well as a normal space, a.k.a normalization. For a comprehensive introduction to fractional orthogonal kernel function refer to Learning with Fractional Orthogonal Kernel Classifiers in Support Vector Machines book.

Getting started:

• Installation
  • Dependencies
  • Conda environment (suggested)
• References
• Credits

How to:

• How to Fit a model
  • Load data-set
  • Initialize a model
  • Kernel Initialization
  • Fit the model and Capture paramaters
  • Inspect the model’s accuracy
  • Save the fitted model to Json file
• How to use a prefitted model
  • Load Dataset
  • Load Json File.
  • Predict using the Json file

Examples:

• Fit data-set using Legendre Kernel
  • Load data-set
  • Initiate kernel
  • Fit the model and Capture paramaters
  • Inspect model’s accuracy
• Fit data-set using Chebyshev Kernel
  • Load data-set
  • Initiate kernel
  • Fit the model and Capture paramaters
  • Inspect model’s accuracy
• Fit data-set using Gegenbauer Kernel
  • Load data-set
  • Initiate kernel
  • Fit the model and Capture paramaters
  • Inspect model’s accuracy
• Fit data-set using Jacobi Kernel
  • Load data-set
  • Initiate kernel
  • Fit the model and Capture paramaters
  • Inspect model’s accuracy

Indices and tables

• Index

• Module Index

• Search Page