- 1、本文档共31页,可阅读全部内容。
- 2、原创力文档(book118)网站文档一经付费(服务费),不意味着购买了该文档的版权,仅供个人/单位学习、研究之用,不得用于商业用途,未经授权,严禁复制、发行、汇编、翻译或者网络传播等,侵权必究。
- 3、本站所有内容均由合作方或网友上传,本站不对文档的完整性、权威性及其观点立场正确性做任何保证或承诺!文档内容仅供研究参考,付费前请自行鉴别。如您付费,意味着您自己接受本站规则且自行承担风险,本站不退款、不进行额外附加服务;查看《如何避免下载的几个坑》。如果您已付费下载过本站文档,您可以点击 这里二次下载。
- 4、如文档侵犯商业秘密、侵犯著作权、侵犯人身权等,请点击“版权申诉”(推荐),也可以打举报电话:400-050-0827(电话支持时间:9:00-18:30)。
查看更多
MACHINE LEARNING
CHAPTER 2: LINEAR MODELS FOR REGRESSION
LINEAR MODELS FOR REGRESSION
1. The concept of regression
2. Maximum Likelihood and Least Square
3. Over‐fitting and Regularization
4. The Bias‐Variance Trade‐off
5. Bayesian Linear Regression
6. Sparse regression
The Bias‐Variance Decomposition
Model too “simple” does not fit the data well
• A biased solution
The Bias‐Variance Decomposition
Model too complex small changes to the
data, solution changes a lot
• A high variance solution
The Bias‐Variance Decomposition
D
Given dataset with N samples, learn
function y (x; D)
0
If you sample a different dataset , you will
D
0
learn different y (x; D )
y (x; D) 0
y (x; D )
The Bias‐Variance Decomposition
y (x; D) 0
y (x; D )
Expected hypothesis: ED [y (x; D)]
Take the average over the
ensemble of data sets.
The Bias‐Variance Decomposition
Bias: difference between what you expected
to learn and the truth
• Measure how well you expected to
represent true solution
• Decreased with more complex model
average learned model truth distribution of input
The Bias‐Variance Decomposition
Variance: difference between what you
expected to learn and what you learn from a
particular dataset
• Measure how sensitive learner is to specific
dataset
• Decreases with simple model
The Bias‐Variance Decomposition
Choice of hypothesis cl
您可能关注的文档
- @[TMM 2010] Real-Time Visual Concept Classification英文文献资料.pdf
- [ICCV 2011] Tasting Families of Features for Image Classification英文文献资料.pdf
- [NIPS 2011] Learning Anchor Planes for Classification英文文献资料.pdf
- [NIPS 2011] Matrix Completion for Multi-label Image Classification英文文献资料.pdf
- [NIPS 2012] Multilabel Classification using Bayesian Compressed Sensing英文文献资料.pdf
- [NIPS 2012] Projection Retrieval for Classification英文文献资料.pdf
- 01_Linear_Models_for_Binary_Classification_21-35英文文献资料.pdf
- 1A - Fudan 2014 Salary Classification Question英文文献资料.pdf
- 2_PRoNTo_Course_Classification英文文献资料.pdf
- 3_06_performing_classification英文文献资料.pdf
最近下载
- 人教PEP版六年级上册Unit 5 What does he do 单元整体教学设计.pdf
- 2021年新生儿感染:抗菌药物预防性与治疗性使用指南解读(全文).pdf
- 《数字图像处理教程》试题库.pdf VIP
- 学校意识形态工作计划.pdf
- 战略模拟软件CESIM全攻略(课堂PPT).ppt
- 中国碳交易政策对可持续经济福利的影响.pptx VIP
- 反三违(典型“三违”)行为清单.docx
- (高清版)B-T 3836.1-2021 爆炸性环境 第1部分:设备 通用要求.pdf VIP
- 人教版高一化学必修一知识点梳理.docx
- 人教pep版英语三年级上册阅读理解专项复习试卷测试题(含答案).doc
文档评论(0)