The demo consists of two parts: source codes and data. In order to reproduce the exact results of experiments in our paper, please use the same simulation data, including testing datasets (MNIST and NORB) and random matrices, which can be provided upon request via email: cwdeng@bit.edu.cn
To use these codes, you can simply unzip all files into the same path and then run the "demo_MNIST.m" and "demo_NORB.m".
The main training function "helm_train()" could be called as follows:
Example:
[TrainingAccuracy, TestingAccuracy, Training_time, Testing_time] = helm_train(train_x, train_y, test_x, test_y, b1, b2, b3, s, C);
% train_x is the training data and train_y is the training label.
% test_x is the training data and test_y is the training label.
% b1, b2 and b3 are the random matrices, they are pre-stored in our demo and can be used by loading the random*.mat.
% C is the L2 penalty of the last layer ELM and s is the scaling factor of the activation function.
Please cite our paper in your publications if it helps your research:
@article{tang2015helm,
author={Tang, Jiexiong and Deng, Chenwei and Huang, Guang-Bin.},
journal={IEEE Transactions on Neural Networks and Learning Systems},
title={Extreme Learning Machine for Multilayer Perceptron},
year={2015},
doi={10.1109/TNNLS.2015.2424995},
ISSN={2162-237X},}
Enjoy, :-).
H-ELM实现人脸识别附matlab代码.zip
版权申诉
68 浏览量
2023-04-30
20:01:15
上传
评论
收藏 5KB ZIP 举报
Matlab科研辅导帮
- 粉丝: 2w+
- 资源: 7666
最新资源
- spice-guest-tools-0.164.4
- 第2次作业-单片机开发板.epro
- 免费进程管理器 ManageEngine-FreeProcessManager
- paraformer的cif模块 文本/alpha 强制对齐脚本
- 北京大学数字普惠金融指数(PKU-DFIIC)2011-2022
- 猪齿鱼Choerodon数智化开发管理平台,提供体系化方法论和协作、测试、DevOps及容器工具,帮助企业拉通需求、设计、开发
- 蚁群算法求解TSP问题
- springboot作业
- funasr 微调数据集制作脚本
- 免费密码过期提示工具 ManageEngine-ADSelfService-Plus-64bit.exe
资源上传下载、课程学习等过程中有任何疑问或建议,欢迎提出宝贵意见哦~我们会及时处理!
点击此处反馈