Giter Site home page Giter Site logo

comments's People

Watchers

 avatar

comments's Issues

regularization | 我啊lkjflaj

https://icream.top/2019/09/21/regularization/#more

有道云笔记 [TOC] cost Function和regularization监督学习的目的就是要最小化cost Function. 但这个过程中往往会产生些错误,如overfit,而正则化的存在是通过调整正则化参数而达到想要的目的 cost Functionlog对数cost Function—logstic regression square loss —– 最小二乘,OLS Hinge

GBDT | 我啊lkjflaj

https://icream.top/2019/07/18/GBDT/

加载不了图片就点有道云笔记 机器学习技法 —— Gradient Boosted Decison Tree==这一章,从第一小节已知延伸到最后的GBDT,其中每一小节都包含了一个不同的aggregation== 1. AdaBoost Decision Tree首先,比对比对上图左右两个,RF和AdaBoost DT,区别就在于base algorithm用的是bagging还是AdaBoost。

SVM与对偶perceptron | 我啊lkjflaj

https://icream.top/2019/10/29/SVM%E4%B8%8E%E5%AF%B9%E5%81%B6perceptron/#more

SVM与对偶Perceptron最近上课讲到SVM,太久没回顾了有点蒙,之前笔记之记在纸上,甚是不便,今日便将其整理upload。 由perceptron引发的这篇文章普通perceptronloss function:梯度更新则有: 对偶perceptron我们可以相对原本的perceptron算法换个思路去思考,因为我们有梯度更新如下,即最终的w = 所有ηyx相加,b = 所有ηy相加

逻辑回归与交叉熵 | 我啊lkjflaj

https://icream.top/2019/09/27/%E9%80%BB%E8%BE%91%E5%9B%9E%E5%BD%92%E4%B8%8E%E4%BA%A4%E5%8F%89%E7%86%B5/

由于逻辑回归是一个伯努利分布,所以逻辑回归本质是交叉熵(作为损失函数)逻辑回归和交叉熵对于MSE顾虑都一样,(y-a)^2,对于神经网络而言,会降低w,b的更新速度,而对于逻辑回归来说,MSE会导致代价函数非凸,存在很多局部最优解。先了解信息熵和相对熵 相对熵—KL散度相对熵又称KL散度,如果我们对于同一个随机变量 x 有两个单独的概率分布 P(x) 和 Q(x),我们可以使用 KL 散度(K

Random Tree | 我啊lkjflaj

https://icream.top/2019/07/16/Random%20Tree/

加载不了图片就点有道云笔记 机器学习技法 —— Random Forest基分类器gt 1. Random Forest Alogorithm回想Bagging, 其可以减少variance,而Decision Tree则是具有很大的variance,那么如果把他们结合起来就有Random Forest同时,这种方法具有相对多的好处,能够并行运算不同的决策树再合起来,RF集成了CART的优点,还

C3 信息增益 | 我啊lkjflaj

https://icream.top/2019/09/04/C3%20%E4%BF%A1%E6%81%AF%E5%A2%9E%E7%9B%8A/#more

看不到图片就点有道云笔记 ==为了对信息(特征?)进行量化处理,然后再选定以什么作为标准,期望(纯度越来越高)信息增益越大== [TOC] 决策树 特征选择“信息熵”是度量样本集合不确定度(纯度)的最常用的指标熵—->不确定性的量度熵越高,不确定性越大越随机,概率低,熵越大熵越大,信息量越大 1. 信息熵信息熵公式: 其中p(xi)代表随机事件X为xi的概率 信息量:(时间发生概率越大,信息

c3 决策树算法比较 | 我啊lkjflaj

https://icream.top/2019/09/04/c3%20%E5%86%B3%E7%AD%96%E6%A0%91%E7%AE%97%E6%B3%95%E6%AF%94%E8%BE%83/

看不到图片就点有道云笔记 [TOC] CART, C4.5, ID3 比较ID3信息增益主导ID3 容易导致一个问题就是加入14个样本,14个都不同种类,算法的结果会是分成14个分支,不容易泛化,由此衍生出C4.5的算法 C4.5信息增益率主导 CARTGini系数主导Gini系数反应的是剩余数据集的纯度,如果Gini系数越小就说明数据集D纯度越高 与信息增益公式类似:==加入特征X后数据不

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.