Algorithm:
 
 
When to select Anonaly detection or Supervised learning? 总的来说guideline是如果positive example (anomaly examples)特别少就用Anamaly detection. 如果数据positive example 越来越多,可以选择从Anomanly detection 切换到 Supervised learning.
 
 
怎么选择feature ?
 
可以先画出feature的分布图,如果不符合高斯分布,就对feature做一些转化,使得更像高斯分布
 
如果一维feature实在不像高斯分布,可以考虑给feature再多加一个维度
 
 
Multivariate Gaussian(Normal) examples
 
The original Gaussian model is a specific multivariate Gaussian model, please see following picture
 
 
 
 
Original vs Multivariate Gaussian
    虽然Multivariate 更全面,但是一般来说Original用的更多一些,因为计算cheeper, 一般在m很大n不大的时候考虑Mutivariate版本
 
 
 

Recommender system

 
  1. content based recommender system (content based 意思是比如我们已经知道每个电影的类型,像爱情片,动作片,。。。)

  人工分析每个电影的内容,得出feature 值,就是下面的romance, action 对应的值. 然后用linear regression 求出 theta.

  

  

  上面讲的是针对一个user的theta 参数,下面这个图是learn 所以的参数

 
  1. colabarative filtering algorithm (for feature learning, for specifically, theta->x->theta->x...)
   

algorithm:
concept of the algorithm of colabarative filtering
cost function:
calculate theta and x simutanously in one cost function
 
 
derivative:
 

  1. low rack matrix factorization ( 不是一种新算法,是collaborating filter algo 的矩阵实现)
    
   

    
Ref:

Coursera, Machine Learning, Anomoly Detection & Recommender system的更多相关文章

  1. Machine Learning No.11: Recommender System

    1. Content based Problem formulation Content Based Recommendations: 2. collaborative filtering algor ...

  2. Coursera, Machine Learning, notes

      Basic theory (i) Supervised learning (parametric/non-parametric algorithms, support vector machine ...

  3. Coursera machine learning 第二周 quiz 答案 Linear Regression with Multiple Variables

    https://www.coursera.org/learn/machine-learning/exam/7pytE/linear-regression-with-multiple-variables ...

  4. 神经网络作业: NN LEARNING Coursera Machine Learning(Andrew Ng) WEEK 5

    在WEEK 5中,作业要求完成通过神经网络(NN)实现多分类的逻辑回归(MULTI-CLASS LOGISTIC REGRESSION)的监督学习(SUOERVISED LEARNING)来识别阿拉伯 ...

  5. 【Coursera - machine learning】 Linear regression with one variable-quiz

    Question 1 Consider the problem of predicting how well a student does in her second year of college/ ...

  6. Coursera, Machine Learning, SVM

    Support Vector Machine (large margin classifiers ) 1. cost function and hypothesis 下面那个紫色线就是SVM 的cos ...

  7. Coursera, Machine Learning, Neural Networks: Representation - week4/5

    Neural Network Motivations 想要拟合一条曲线,在feature 很多的情况下,feature的组合也很多,在现实中不适用,比如在computer vision问题中featu ...

  8. Coursera machine learning 第二周 编程作业 Linear Regression

    必做: [*] warmUpExercise.m - Simple example function in Octave/MATLAB[*] plotData.m - Function to disp ...

  9. Coursera machine learning 第二周 quiz 答案 Octave/Matlab Tutorial

    https://www.coursera.org/learn/machine-learning/exam/dbM1J/octave-matlab-tutorial Octave Tutorial 5  ...

随机推荐

  1. ArcGIS for qml -设置地图和视域中心

    源码:https://github.com/sueRimn/ArcGIS-for-qml-demos 作者: 狐狸家的鱼 Github: 八至 版权声明:如需转载请获取授权和联系作者 ArcGIS R ...

  2. WIndows下将文件夹映射为磁盘

    subst 盘符 文件夹路径 [/d] 映射 将e:\work映射为z:盘,使用subst z: e:\work 取消映射 取消z盘映射,使用subst z: /d 参考资料:http://mp.we ...

  3. A1009. Product of Polynomials

    This time, you are supposed to find A*B where A and B are two polynomials. Input Specification: Each ...

  4. tensorflow不同版本安装与升级/降级

    https://blog.csdn.net/junmuzi/article/details/78357371 首先,可以安装一个anaconda. 然后使用python的pip可以安装特定版本的ten ...

  5. python(字符串、列表、字典、元组、集合)的常用内置方法

    一.字符串: lis='my name is maple' res=lis.count('m')#计算字符串内相同字符的个数 print(res) lis='my name is maple' res ...

  6. GWAS研究可利用的数据库(持续更新)

    1.列表包括数据库名称.表型.是否能下载到基因型(genotype).是否能下载到GWAS结果文件(P值.效应值.SNP位点).目前收集到的有如下: 参考到这些数据库的文献:Genome-wide a ...

  7. POJ 2823 Sliding Window (单调队列)

    单调队列 加了读入挂比不加更慢.... 而且这份代码要交c++ 有大神G++跑了700ms..... orzorzorz #include<iostream> #include<cs ...

  8. 做web开发需要学习哪些技术--基础篇

    做一个web网站,包含哪些技术,自己需要学习哪些技术 自己想到哪里就写到哪里 -- 给自己做的一个记录 1: 页面的展示, 一个web的开发语言  1.1 一个web的开发语言需要注意哪方面,才能符合 ...

  9. linux ssh 报错failed - POSSIBLE BREAK-IN ATTEMPT

    linux ssh 报错failed - POSSIBLE BREAK-IN ATTEMPT 问题故障: 今天在新租的虚拟机上,发现ssh登陆机器的时候报错,如下: [root@pictures_ne ...

  10. Storm 消息分发策略

    1.Shuffle Grouping:随机分组,随机派发stream里面的tuple,保证每个bolt接收到的tuple数目相同.2.Fields Grouping:按字段分组,比如按userid来分 ...