Felix_夜雨-CSDN博客

原创 Machine Learning Overview

Journal: http://jmlr.csail.mit.edu/papers/Blog: http://hunch.net/Surveys: http://www.mlsurveys.com/Tutorials: http://www.autonlab.org/tutorials/Lectures: http://www.cs.cmu.edu/~tom/10701_sp11/lectures

2013-06-03 12:15:06 1367 1

原创 Study notes for Sampling

Study notes for Sampling

2013-09-12 21:41:45 934

原创 Study notes for Sparse Coding

Study notes for Sparse coding and comments.

2013-09-11 23:10:52 2317

原创 Study notes for B-tree and R-tree

Study notes for B-tree and its variants, and R-tree.

2013-09-10 22:50:28 1630

原创 Study notes for Metric Trees

Study notes for metric trees, including introductions to metric space, metric trees; and spatial indexing trees such as M-tree.

2013-09-06 11:59:46 1636

原创 Study notes for Continuous Probability Distributions

This article describes the basics of probability and the most commonly used continuous probability distributions

2013-07-09 11:39:12 1342

原创 Study notes for Discrete Probability Distribution

This article describes the basics of probability and the most commonly used discrete probability distributions.

2013-07-08 17:07:59 1189

原创 Study notes for Clustering and K-means

1. Clustering AnalysisClustering is the process of grouping a set of (unlabeled) data objects into multiple groups or clusters such that objects within a cluster have high similarity, but are very

2013-07-01 15:05:19 1552

原创 Research notes for Transfer Learning

IntroductionTransfer learning (a.k.a, cross-domain learning or domain adaptation) is an emerging research topic in computer vision and recommender systems, etc. It is well-known that the feature d

2013-06-30 14:57:16 1283

原创 Study notes for Gaussian Mixture Model

1. Mixture ModelA typical k-dimensional mixture model is a hierarchical model consisting of:There are N observations, each observation is a mixture of K components. Each component belongs to t

2013-06-23 17:37:14 2127

原创 Study notes for Latent Dirichlet Allocation

1. Topic ModelsTopic models are based upon the idea that documents are mixtures of topics, where a topic is a probabilistic distribution over words. A topic model is a generative model for docum

2013-06-22 23:06:48 2997

原创 Study notes for Anomaly Detection

IntroductionAssume we have a data set including vast normal examples and only small anomalies. To detect these anomalies, we need to train a probabilistic model to determine whether a given exampl

2013-06-19 21:38:31 1666

原创 Study notes for Expectation Maximum Algorithm

1. IntroductionThe EM algorithm is an efficient iterative procedure to compute the maximum likelihood (ML) estimate in the presence of missing or hidden data (variables). It intends to estimate th

2013-06-15 23:11:15 2493

原创 Study notes for Non-negative Matrix Factorization

This article gives an overview of nonnegative matrix factorization used in machine learning and recommender systems.

2013-06-13 22:50:16 5693 1

原创 Study notes for Principal Component Analysis

Motivations for Dimensionality ReductionData ComparisonSpeed up algorithms. By reducing a large number (e.g. 10, 000) of dimensionality of feature space, a learning algorithm may be too slow t

2013-06-11 22:44:20 1339

转载 Study notes for Support Vector Machine (1)

Note: originally posted in http://www.holehouse.org/mlclass/12_Support_Vector_Machines.htmlSupport Vector Machine (SVM) - Optimization objectiveSo far, we've seen a range of different algo

2013-06-08 12:31:38 920

原创 Study notes for Feature Engineering

Study notes for feature engineering including selection and extraction of features. Feature engineering is key to machine learning.

2013-06-05 15:54:14 2652

原创 English notes for Disease

Notes: most contents may be collected from multiple sites, only for study purpose.一般病情He feels headache, nausea and vomiting. (他觉得头痛、恶心和想吐。)He is under the weather. (他不舒服，生病了。)He began to feel

2013-06-02 17:29:24 817

转载应用机器学习的建议(Advice for applying machine learning)

转载自：Coursera公开课笔记: 斯坦福大学机器学习第十课“应用机器学习的建议(Advice for applying machine learning)”其它笔记：Stanford机器学习---第六讲. 怎样选择机器学习方法、系统 (Note: Rachel Zhang explains in detail the concepts of bias and variance. Q

2013-05-29 13:56:23 1651

转载机器学习系统设计(Machine learning system design)

转载自：Coursera公开课笔记: 斯坦福大学机器学习第十一课“机器学习系统设计(Machine learning system design)”其它笔记：Stanford机器学习---第七讲. 机器学习系统设计斯坦福大学机器学习斯坦福大学机器学习第十一课“机器学习系统设计(Machine learning system design)””学习笔记，本次课程主要包括5部分：

2013-05-29 13:45:56 3181

转载逻辑回归(Logistic Regression)

转载自：Coursera公开课笔记: 斯坦福大学机器学习第六课“逻辑回归(Logistic Regression)”斯坦福大学机器学习第六课"逻辑回归“学习笔记，本次课程主要包括7部分：1) Classification(分类)2) Hypothesis Representation3) Decision boundary(决策边界)4) Cost function

2013-05-29 13:40:41 1137

原创 Be happier and smarter in the life

Notes: most contents may be collected from multiple sites, only for study purpose.Don't compare yourself to others. 不要把自己和别人比较Everybody, and I mean everybody, starts out in a different place and

2013-05-23 21:17:38 761

原创 English notes for the Female

Notes: most contents may be collected from multiple sites, only for study purpose.old hen 老婆子俗语, 男性用语, 专用于中年以上的女性.除开玩笑的场合外, 女性不太欢迎这一称谓.因为 old hen 的字面意思是“老母鸡”.这一表达的来源是: 从男人们的眼光来看, 母鸡来回乱跑, 凑在一起唧唧喳

2013-05-23 21:13:12 701

原创 English notes for Describing People

Notes: most contents may be collected from multiple sites, only for study purpose.a wise guy 骄傲自大的人.这里的wise并不表示“聪明”, 相反它含有讽刺的意味. Alic's all right, I guess, but sometimes he's such a wise guy. 据我

2013-05-23 21:11:11 1119

原创 English notes for Writing Emails

Notes: most contents may be collected from multiple sites, only for study purpose.I am writing to confirm /enquire/inform you...我写信时要确认/询问/通知你...I am writing to follow up on our earlier decision o

2013-05-23 21:08:56 750

原创 English notes for an Interview

Notes: most contents may be collected from multiple sites, only for study purpose.The starting salary I require is HK$6,500 per month. 待遇方面, 本人希望月薪六千五百港币起薪. I require a salary of $4,500 a month

2013-05-23 21:06:49 544

原创 English notes for the Office

Notes: most contents may be collected from multiple sites, only for study purpose.What are you up to? 你正在做什么?假设你在办公室想跟同事沟通下工作情况, 又怕耽误对方的工作, 就可以先问同事"In the middle of something?"(你在忙吗?)如果对方回答"Kind

2013-05-23 21:04:23 574

原创 English notes for Having a Meeting

Notes: most contents may be collected from multiple sites, only for study purpose.Since everyone's here, shall we get started? 大家到齐了, 我们开始吧. I'd like to call the meeting to order. 我宣布会议开始. Today,

2013-05-23 21:02:24 632

原创 English notes for Getting Angry

Notes: most contents may be collected from multiple sites, only for study purpose.Hey!wise up! 放聪明点好吗？当别人做了蠢事时,你可以说,“Don't be stupid！”或“Don't be silly.”但这是非常不礼貌的说法。客气一点的说法就是：Wise up! 你也可以用尖酸刻薄的语气说

2013-05-23 21:00:43 671

原创 English notes for Appreciating Others

Notes: most contents may be collected from multiple sites, only for study purpose.You look great today.（你今天看上去很棒.）You did a good job. （你干得非常好.） We're so proud of you.（我们十分为你骄傲.）(最高级的表扬！) I'm very

2013-05-23 20:58:25 520

原创 English notes for Making a Call

Notes: most contents may be collected from multiple sites, only for study purpose.I can't get through. 电话打不通.还能这样说: The line isn't free. The number is busy now. 应用解析: get back 回来; 取回; 报复; get be

2013-05-23 20:52:58 526

原创 English notes for Time

Notes: most contents may be collected from multiple sites, only for study purpose.The homework is due two weeks from Monday.这个作业在二个星期后的星期一交。在英文中要提到下星期一可以用 next Monday，下下星期一偷懒的话可以说 next next Mod

2013-05-23 20:47:05 664

原创 English notes for Commonly Used Phrases

Notes: most contents may be collected from multiple sites, only for study purpose.解决: Solve, deal with, cope with, handle, resolve, address, tackle 损害: Damage, hurt, injure, harm, impair, undermin

2013-05-23 20:42:42 692

原创 Study notes for Support Vector Machines (2)

It is a method for the classification of both linear and nonlinear data, i.e., linearly separable and nonlinearly separable data. In addition, SVMs can be used for numeric prediction. Put simply, SV

2013-05-23 09:00:36 850

原创 Study notes for Backpropagation

It can be used as a linear and nonlinear classifier as well as a non-linear regression (i.e. for numeric prediction). Multilayer feed-forward networks, given enough hidden units and enough training sa

2013-05-22 23:02:54 657

原创 Python Startup

1. Python Modules InstallationPython shell: use "python py.setup install" in windows prompt; or "py.setup install" in python prompt. iPython shell:tab: complement your codescd: display or cha

2013-05-22 16:04:19 1335

原创 Study notes for Decision Trees

Decision Trees are biased to prefer classifiers that can be expressed as trees. Good tutorials are given by Andrew Moore: Decision Tree and Information Gain. The most well-known algorithms are ID3,

2013-05-21 16:04:23 843

The Rain in the Night