T-Obuchi's repositories
SLRpackage_AcceleratedCV_matlab
Sparse linear regression package with accelerated cross-validation. L_1, SCAD, MCP penalties are covered. The algorithm for optimization is cyclic coordinate descent.
AMPR_lasso_python
Bootstrap resampling is used to estimate confidence interval of variables in Lasso (some famous methods are bolasso and stability selection). This MATLAB package performs this in an efficient manner by conducting the resampling in a semi-analytic manner, enabling to avoid numerical resampling. MATLAB version is available: https://github.com/T-Obuchi/AMPR_lasso_matlab
AcceleratedCVonMLR_python
This Python package enables to efficiently compute leave-one-out cross validation error for multinomial logistic regression with elastic net (L1 and L2) penalty. The computation is based on an analytical approximation, which enables to avoid re-optimization and to reduce much computational time. MATLAB version: https://github.com/T-Obuchi/AcceleratedCVonMLR_matlab
AMPR_lasso_matlab
Bootstrap resampling is used to estimate confidence interval of variables in Lasso (some famous methods are bolasso and stability selection). This MATLAB package performs this in an efficient manner by conducting the resampling in a semi-analytic manner, enabling to avoid numerical resampling. Python translation is available: https://github.com/T-Obuchi/AMPR_lasso_python
AcceleratedCVon2DTVLR
This MATLAB package enables to efficiently compute leave-one-out cross validation error for linear regression with two regularization terms: L_1 and total-variation. The computation is based on an analytical approximation, which enables to avoid re-optimization and to reduce much computational time.
AcceleratedCVonMLR_matlab
This MATLAB package enables to efficiently compute leave-one-out cross validation error for multinomial logistic regression with elastic net (L1 and L2) penalty. The computation is based on an analytical approximation, which enables to avoid re-optimization and to reduce much computational time. Python version: https://github.com/T-Obuchi/AcceleratedCVonMLR_python
graphBIX
Graph clustering by Bayesian inference with cross-validation model assessment