没有合适的资源?快使用搜索试试~ 我知道了~
本文档详细阐释了将深度学习相关理论成果应用于SVM使SVM机器学习效果性能更加优良
资源推荐
资源详情
资源评论
Machine Learning
I
Machine learning algorithms are very useful for many
applications (Classification, Regression, Adaptive Control)
I
They use datasets or experiences to fit a model
I
Example applications are:
I
Object Recognition
I
Face Recognition
I
Handwritten Digit Recognition
I
fMRI-Scan Classification
I
Medical Diagnosis
I
Document Classification
I
Many Applications in Bio-informatics and Chemistry
Marco A. Wiering 3/25
Limitations of Support Vector Machines
I
Support Vector Machines (SVM) often outperform other
machine learning methods
I
However, the standard SVM has a single adjustable layer
of weights
I
Instead of using such “shallow models”, deep architectures
can be better alternatives
I
SVMs use a-priori chosen kernel functions to compute
similarities between input vectors
I
A problem is that the choice of kernel function is important,
but kernel functions are not very flexible
I
Therefore we propose the deep SVM (DSVM)
I
The DSVM contains multiple layers of SVMs
Marco A. Wiering 4/25
Support Vector Regression
I
The objective function of the SVM is based on structural
risk minimization theory developed by Vapnik in the 1960s
I
Goal: find g(x) most suitable to the data, e.g. for
regression, -insensitive (Hinge) loss function:
I
|y
i
− g(x
i
)| ≤ ε
I
But also generalize well!
I
g(x) as flat as possible ⇒ ||w|| as small as possible
I
Yields a convex optimization problem
Marco A. Wiering 5/25
剩余24页未读,继续阅读
资源评论
finishspan
- 粉丝: 1
- 资源: 1
上传资源 快速赚钱
- 我的内容管理 展开
- 我的资源 快来上传第一个资源
- 我的收益 登录查看自己的收益
- 我的积分 登录查看自己的积分
- 我的C币 登录后查看C币余额
- 我的收藏
- 我的下载
- 下载帮助
最新资源
- 基于keras+fasterRCNN,在VOC格式的口罩数据集上训练,检测人群中有无戴口罩python源码+模型
- push_version
- 软件自制图像批量压缩工具
- 基于深度学习的抗梯度噪声的缺陷检测器python源码+文档说明+模型的预训练
- 基于python+pytorch+mysql实现停车场车牌识别管理系统源码+文档说明
- 基于QT+MySQl+OpenCV车牌识别搭建停车场管理系统C++源码+文档说明+界面展示
- 基于深度学习的停车场收费系统-车牌识别模块python源码+文档说明+博客教学
- 空白.pages
- 基于Java+Springboot+vue的智能停车场管理系统(源代码+数据库+9000字论文) 本项目前后端不分离+部署教程
- 基于SSM写的停车场管理系统,加入了车牌识别和数据分析+源码+文档说明
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
安全验证
文档复制为VIP权益,开通VIP直接复制
信息提交成功