TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Guide to Linear Regression

67 点作者 alexhwoods将近 10 年前

5 条评论

gpcz将近 10 年前
A related really cool thing I learned in grad school is that you can implement Deming regression (<a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Deming_regression" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Deming_regression</a>) by storing the moments (outer product sum and point sum) of the training points and then finding the dominant eigenvector of the outer product sum using a singular value decomposition. Since you can approximate the directional part of a 2x2 SVD with atan2(), it effectively becomes a O(1) operation to add or remove training points.<p>(specifics here: <a href="https:&#x2F;&#x2F;april.eecs.umich.edu&#x2F;courses&#x2F;eecs568_f12&#x2F;linefitting.pdf" rel="nofollow">https:&#x2F;&#x2F;april.eecs.umich.edu&#x2F;courses&#x2F;eecs568_f12&#x2F;linefitting...</a> )
评论 #9927322 未加载
评论 #9930286 未加载
bluusteel将近 10 年前
I thought closed form expressions existed for linear regression. Why is gradient descent needed?
评论 #9927661 未加载
评论 #9927612 未加载
cstuder将近 10 年前
This is a nice guide, I&#x27;ve also enjoyed playing around with the Ipython Notebook here: <a href="http:&#x2F;&#x2F;nbviewer.ipython.org&#x2F;github&#x2F;justmarkham&#x2F;DAT4&#x2F;blob&#x2F;master&#x2F;notebooks&#x2F;08_linear_regression.ipynb" rel="nofollow">http:&#x2F;&#x2F;nbviewer.ipython.org&#x2F;github&#x2F;justmarkham&#x2F;DAT4&#x2F;blob&#x2F;mas...</a><p>The one question which remains: Is there a more intuitive guide which helps you with deciding which features you need to choose in order to get a good regression?
评论 #9930284 未加载
kastnerkyle将近 10 年前
For anyone who is interested in a python version with inferior plots I have a notebook here:<p><a href="http:&#x2F;&#x2F;kastnerkyle.github.io&#x2F;posts&#x2F;linear-regression&#x2F;" rel="nofollow">http:&#x2F;&#x2F;kastnerkyle.github.io&#x2F;posts&#x2F;linear-regression&#x2F;</a><p>Closed form is at the bottom, basically invalidating the whole rest of my post.
npx将近 10 年前
Gonna use this as the title for my autobiography! :)