matlab连乘的代码-OSVR:序数支持向量回归

  • G1_223222
    了解作者
  • 8.3MB
    文件大小
  • zip
    文件格式
  • 0
    收藏次数
  • VIP专享
    资源类型
  • 0
    下载次数
  • 2022-05-14 10:50
    上传日期
matlab连乘的代码OSVR 序数支持向量回归(OSVR)是一种通用回归模型,它以数据样本及其成对序数关系为输入,并在最大边距框架下输出从数据中学习到的模型参数。 不需要每个单独样本的标签(回归响应)。 OSVR 可以从无监督(无标签)到完全监督(所有样本上都有标签)。 特别是,OSVR 在弱监督下非常有用,其中仅提供选定关键样本的标签。 但是,在不同的监督设置下,顺序信息应始终可用。 当前版本 1.0 仅支持线性回归模型。 如何使用 该存储库提供了在 Matlab 中 OSVR 的实现。 核心优化问题是使用定制的乘法器交替方向法 (ADMM) 解决的。 然而,'admm' 函数的实现足够通用,可以处理标准形式的优化问题。 有关更多详细信息,请参阅相关出版物。 要查看示例,请运行“main”脚本。 提供的样本数据对应于 UNBC-McMaster 肩痛数据集的一个测试折叠 [1]。 如果您需要更多数据进行比较,请联系作者。 相关刊物 如果您在研究中使用该代码,请考虑引用以下出版物 赵瑞、甘泉、王尚飞、季强。 “使用序数信息进行面部表情强度估计。” IEEE 计算机视觉和模式识别会议。
OSVR-master.zip
  • OSVR-master
  • constructParams.m
    3.3KB
  • data.mat
    8.3MB
  • README.md
    2.1KB
  • ICC.m
    2.8KB
  • admm.m
    6.1KB
  • main.m
    2.4KB
内容介绍
## OSVR Ordinal Support Vector Regression (OSVR) is a general purpose regression model that takes data samples as well as their pairwise ordinal relation as input and output the model parameters learned from data under the max-margin framework. Label (regression response) of each individual sample is not required. OSVR can adapt from no supervision (no labels) to full supervision (labels on all samples). In particular, OSVR can be very useful under weak supervision, where only labels on selected key samples are provided. However, ordinal information should always be available under different supervision settings. Current version 1.0 only supports linear regression model. ## How to use This repository provides an implementation of OSVR in Matlab. The core optimization problem is solved using customized Alternating Direction Method of Multipliers (ADMM). However, the 'admm' function is implemented general enough to handle an optimization problem with standard form. See related publication for more details. To see an example, run 'main' script. The sample data provided correspond to one testing fold of UNBC-McMaster shoulder pain dataset [1]. Please contact author if you need more data for comparison purpose. ## Related publication If you use the code in your research, please consider citing following publication Rui Zhao, Quan Gan, Shangfei Wang and Qiang Ji. "Facial Expression Intensity Estimation Using Ordinal Information." IEEE Conference on Computer Vision and Pattern Recognition. 2016. [[PDF](http://www.cv-foundation.org/openaccess/content_cvpr_2016/papers/Zhao_Facial_Expression_Intensity_CVPR_2016_paper.pdf)] [[Bib](http://www.cv-foundation.org/openaccess/content_cvpr_2016/html/Zhao_Facial_Expression_Intensity_CVPR_2016_paper.html)] [[Supplement](http://homepages.rpi.edu/~zhaor/document/Zhao2016_supp.pdf)] ### Reference [1] P. Lucey, J. F. Cohn, K. M. Prkachin, P. E. Solomon, and I. Matthews. Painful data: The unbc-mcmaster shoulder pain expression archive database. In FG, pages 57–64. IEEE, 2011. ## License Conditions Copyright (C) 2016 Rui Zhao Distibution code version 1.0 - 06/25/2016. This code is for research purpose only.
评论
    相关推荐