Document of Linux

Install


Using the conda command to install SBSM onto your computer: 

conda install -c malab -c conda-forge sbsm

Usage

Sbsm is a command-line tool that can be run by simply typing sbsm, results as shown in the figure below



The common commands for training using sbsm are shown below, which will result in a trained model file results.model.

sbsm -t train_samples.fasta train_labels.txt

The common commands for testing using sbsm are shown below, which will result in a predicted label file predicted_label.txt.

sbsm -p results.model test_samples.fasta

Options


The sbsm tool offers a variety of options to customize the training and prediction processes:

  • -c, --c: Penalty coefficients for support vector machines. Default value: [64]
  • -a, --alpha: Control parameter for kernel parameterization. Default value: [-1]
  • -m, --match: Match score. Default value: [1]
  • -x, --mismatch: Mismatch penalty. Default value: [-1]
  • -g, --gap: Gap penalty. Default value: [-2]
  • -k, --kneighbors: In HCKDM, k represents the number of neighboring kernels chosen for local kernel selection. Default value: [15]
  • -l, --lambda: Ratio parameter of the global kernel alignment in both global and local kernel alignments. Default value: [0.9]
  • -n1, --nu1: Regularization parameter for the Laplacian graph regularization term. Default value: [0.01]
  • -n2, --nu2: Regularization parameter for the L2 regularization term. Default value: [0.01]
  • -h, --help: Print the help usage information.
  • -v, --version: Show the version number of sbsm.

Manual

Download the manual of SBSM.


Download manual

Malab.