APGD v.0.1.0 (Python Version)

Python version of the Accelerated Proximal Gradient Descent (APGD) algorithm is to solve the penalized regression models, including

• HuberNet: Huber loss function along with Network-based penalty function;
• HuberLasso: Huber loss function along with Lasso penalty function;
• HuberENET: Huber loss function along with Elastic Net penalty function;
• ENET: Mean square error loss function along with Elastic Net penalty function;
• Lasso: Mean square error loss function along with Lasso penalty function;
• Net: Mean square error loss function along with Network-based penalty function.

We also have R version, please see the following link for the guideline of R version https://github.com/xueweic/APGD.

Reference

Xuewei Cao⁺, Ling Zhang⁺, Kui Zhang, Sanzhen Liu, Qiuying Sha*, Hairong Wei*. HuberNet function for interfering target genes of regulatory genes using high-throughput gene expression data.

_{^{+ These authors have contributed equally to this work}}

Any questions? lingzhan_AT_mtu_DOT_edu, xueweic_AT_mtu_DOT_edu

Installation

Please use Python Version 3

Step 1. Download 'requrirements.txt' file for installing the requirements packages in your terminal:

pip install -r requirements.txt

Step 2. Simple install APGD package by runing command:

pip install APGD

Step 3. Test your APGD in python:

import APGD

If there is no Error, you have installed APGD package successfully!

Step 4. Play with function in APGD referred by APGD_Guide.

Functions

Please refer APGD_Guide.