Boostable-CSDN博客

原创 Jordan Lecture Note-12: Kernel典型相关分析(Kernel Canonical Correlation Analysis, KCCA).

Kernel典型相关分析（一）KCCA 同样，我们可以引入Kernel函数，通过非线性的坐标变换达到之前CCA所寻求的目标。首先，假设映射$\Phi_X: x\rightarrow \Phi_X(x), \Phi_Y: y\rightarrow \Phi_Y(y)$，记$\mathbf{\Phi_X}=(\Phi_X(x_1),\Phi_X(x_2),\cdots,\Phi_X(x_

2014-03-10 12:19:09 1765

原创王道论坛研究生机试练习赛（二）

昨天在微博上看到王道论坛说晚上有机试练习赛（http://ac.jobdu.com/contest.php?cid=1053），想来反正也没什么事，何不参加一下。于是晚上吃完饭回到实验室，等着七点开始做题。练习赛结束了，虽然自己全部AC，但感觉时间花的有点长，主要卡在第三道求公约数个数上。看来自己的数论真的弱爆了。现将自己对这四道解题的看法归纳一下，总的来说题目不算很难，前两道超级水，第三道考素数

2014-03-10 12:19:07 658

原创 Jordan Lecture Note-11: 典型相关分析(Canonical Correlation Analysis, CCA).

典型相关分析（一）引入典型相关分析（Canonical Correlation Analysis）是研究两组变量之间相关关系的一种多元统计方法。他能够揭示出两组变量之间的内在联系。我们知道，在一元统计分析中，用相关系数来衡量两个随机变量的线性相关关系，用复相关系数研究一个随机变量与多个随机变量的线性相关关系。然而，这些方法均无法用于研究两组变量之间的相关关系，于是提

2014-03-10 12:19:04 732

原创 PAT 1021

1021. Deepest Root (25)A graph which is connected and acyclic can be considered a tree. The height of the tree depends on the selected root. Now you are supposed to find the root that results in a

2014-03-10 12:19:02 558

原创 Jordan Lecture Note-10: Kernel Principal Components Analysis (KPCA).

Kernel Principal Components Analysis PCA实际上就是对原坐标进行正交变换，使得变换后的坐标之间相互无关，并且尽可能保留多的信息。但PCA所做的是线性变换，对于某些数据可能需要通过非线性变换，比如在二维空间下对如下数据进行处理。如果还是采用最初的PCA，则得到的主成分是$z_1,z_2$，而这里的$z_1,z_2$都包含了大量的信息，故无法去掉任何一个

2014-03-10 12:19:00 710

原创 PAT 1020

1020. Tree Traversals (25)Suppose that all the keys in a binary tree are distinct positive integers. Given the postorder and inorder traversal sequences, you are supposed to output the level order

2014-03-10 12:18:58 484

原创 PAT 1019

1019. General Palindromic Number (20)A number that will be the same when it is written forwards or backwards is known as a Palindromic Number. For example, 1234321 is a palindromic number. All sin

2014-03-10 12:18:56 480

原创 PAT 1018

1018. Public Bike Management (30)There is a public bike service in Hangzhou City which provides great convenience to the tourists from all over the world. One may rent a bike at any station and re

2014-03-10 12:18:54 543 1

原创 Jordan Lecture Note-9: Principal Components Analysis (PCA).

Principal Components Analysis （一）引入PCA 当我们对某个系统或指标进行研究时往往会发现，影响这些系统和指标的因素或变量的数量非常的多。多变量无疑会为科学研究带来丰富的信息，但也会在一定程度上增加工作的难度，而通常变量之间又具有一定的相关性，这又增加了问题分析的复杂度。如果分别分析每个变量，那么分析又不够综合，而盲目的减少变量又会损失很多有用的信息。因而

2014-03-10 12:18:52 541

原创 PAT 1017

1017. Queueing at Bank (25)Suppose a bank has K windows open for service. There is a yellow line in front of the windows which devides the waiting area into two parts. All the customers have to wait

2014-03-10 12:18:50 511

原创 PAT 1016

1016. Phone Bills (25) A long-distance telephone company charges its customers by the following rules: Making a long-distance call costs a certain amount per minute, depending on the

2014-03-10 12:18:48 524

原创 PAT 1015

1015. Reversible Primes (20) A reversible prime in any number system is a prime whose "reverse" in that number system is also a prime. For example in the decimal system 73 is a reversi

2014-03-10 12:18:46 518

原创 Jordan Lecture Note-8: The Sequential Minimal Optimization Algorithm (SMO).

The Sequential Minimal Optimization Algorithm (SMO)本文主要介绍用于解决SVM对偶模型的算法，它于1998年由John Platt在论文“Sequential Minimal Optimization:A Fast Algorithm for Training Support Vector Machines”中提出的。这篇笔记还参考了某篇博客，

2014-03-10 12:18:44 574

原创 Jordan Lecture Note-7: Soft Margin SVM

Soft Margin SVM （1）Recall 之前分析到SVM的模型为：\begin{align}\mathop{\min}&\quad \frac{1}{2}w^\prime w\nonumber\\\mathop{s.t}&\quad y_i(x_i^\prime w+b)\geq 1, i=1,2,\cdots,m\label{model:SVM}\end{align} 利用L

2014-03-10 12:18:41 637

原创 PAT 1014

1014. Waiting in Line (30) Suppose a bank has N windows open for service. There is a yellow line in front of the windows which devides the waiting area into two parts. The rules for

2014-03-10 12:18:39 606

原创 Jordan Lecture Note-6: The Solutions of Nonlinear Equation.

The Solutions of Nonlinear Equation 本文主要介绍几种用于解非线性方程$f(x)=0$的一些方法。（1） Bisection Method. 算法：step 1: 初始化$a,b(b>a)$，使$f(a),f(b)$异号。step 2: while (停止条件不满足) $p=a+\frac{b-a}{2}$；

2014-03-10 12:18:37 607

原创 PAT 1013

1013. Battle Over Cities (25) It is vitally important to have all the cities connected by highways in a war. If a city is occupied by the enemy, all the highways from/toward that city

2014-03-10 12:18:35 511

原创 Jordan Lecture Note-5: Kernels

Kernels 我们首先来回顾kernel函数的定义：一个函数$K(x,y)$为kernel函数当且仅当对$\forall g, \int K(x,y)g(x)g(y)dxdy\geq 0$成立。另外，根据Mercer's theorem，存在一个映射$\Phi$使$K(x,y)=\langle \Phi(x),\Phi(y)\rangle$，并且对任意有限的点，kernel矩阵是半正定的。

2014-03-10 12:18:33 619

原创 Jordan Lecture Note-4: Linear & Ridge Regression

Linear & Ridge Regression 对于$n$个数据$\{(x_1,y_1),(x_2,y_2),\cdots,(x_n,y_n)\},x_i\in\mathbb{R}^d,y_i\in\mathbb{R}$。我们采用以下矩阵来记上述数据：\begin{equation}\mathbf{X}=\left[\begin{array}& x_1^\prime\\ x_2^\pri

2014-03-10 12:18:31 523

原创 PAT 1012

1012. The Best Rank (25) To evaluate the performance of our first year CS majored students, we consider their grades of three courses only: C - C Programming Language, M - Mathematics

2014-03-10 12:18:29 451

原创 PAT 1011

1011. World Cup Betting (20) With the 2010 FIFA World Cup running, football fans the world over were becoming increasingly excited as the best players from the best teams doing battles

2014-03-10 12:18:27 510

原创 PAT 1010

1010. Radix (25) Given a pair of positive integers, for example, 6 and 110, can this equation 6 = 110 be true? The answer is "yes", if 6 is a decimal number and 110 is a binary number

2014-03-10 12:18:25 476

原创 Jordan Lecture Note-3: 梯度投影法

Jordan Lecture Note-3：梯度投影法在这一节，我们介绍如何用梯度投影法来解如下的优化问题：\begin{align} \mathop{\min}&\quad f(x)\nonumber\\\mathop{s.t.}&\quad \mathbf{A}_1 x\leq b_1\nonumber\\&\quad \mathbf{A}_2x= b_2\label{equ:orig

2014-03-10 12:18:23 6089

原创 PAT 1009

1009. Product of Polynomials (25) This time, you are supposed to find A*B where A and B are two polynomials. Input Specification: Each input file contains one test case. Each case oc

2014-03-10 12:18:20 457

原创 PAT 1008

1008. Elevator (20) The highest building in our city has only one elevator. A request list is made up with N positive numbers. The numbers denote at which floors the elevator will stop

2014-03-10 12:18:18 428

原创 Jordan Lecture Note-2: Maximal Margin Classifier

Maximal Margin Classifier Logistic Regression 与 SVM 思路的不同点：logistic regression强调所有点尽可能远离中间的那条分割线，而SVM则强调最靠近分割线的点于分割线的距离仅可能的远。定义间隔函数：$\hat{r}^{(i)}=y^{(i)}(w^\prime x^{(i)}+b)$。当$y^{(i)}=1$时，$w^\pri

2014-03-10 12:18:16 746

原创 PAT 1007

1007. Maximum Subsequence Sum (25) Given a sequence of K integers { N1, N2, ..., NK }. A continuous subsequence is defined to be { Ni, Ni+1, ..., Nj } where 1 Maximum Subsequence is th

2014-03-10 12:18:14 494

原创 PAT 1006

1006. Sign In and Sign Out (25) At the beginning of every day, the first person who signs in the computer room will unlock the door, and the last one who signs out will lock the door.

2014-03-10 12:18:12 411

原创 PAT 1005

1005. Spell It Right (20) Given a non-negative integer N, your task is to compute the sum of all the digits of N, and output every digit of the sum in English. Input Specification:

2014-03-10 12:18:10 529

原创 PAT 1004

1004. Counting Leaves (30) A family hierarchy is usually presented by a pedigree tree. Your job is to count those family members who have no child. Input Each input file contains one

2014-03-10 12:18:08 500

原创 Jordan Lecture Note-1: Introduction

Jordan Lecture Note-1： Introduction 第一部分要整理的是Jordan的讲义，这份讲义是我刚进实验室时我们老师给我的第一个任务，要求我把讲义上的知识扩充出去，然后每周都要讲给他听。如果有需要这份讲义的话，请留言，我会用邮件发给你。首先，我来说说机器学习这个东西。刚进实验室，我根本连什么是机器学习都不知道，听到这个名词后的第一反应是机器人，心想估计是

2014-03-10 12:18:06 827 1

原创 PAT 1003

1003. Emergency (25) As an emergency rescue team leader of a city, you are given a special map of your country. The map shows several scattered cities connected by some roads. Amount

2014-03-10 12:18:04 442

原创 PAT 1002

1002. A+B for Polynomials (25) This time, you are supposed to find A+B where A and B are two polynomials. Input Each input file contains one test case. Each case occupies 2 lines, and

2014-03-10 12:18:02 477

原创 PAT 1001

Calculate a + b and output the sum in standard format -- that is, the digits must be separated into groups of three by commas (unless there are less than four digits). Input Each input file conta

2014-03-10 12:18:00 415

原创写在最前

一直以来都想开通一个博客来记录自己上研究生以来的学习过程，可每次都因为各种事情而耽搁着，总是以当前的事情为借口将其推迟，归根结底还是自己太懒了。考上研究生后，我莫名其妙的选择了机器学习这个方向，还好接触后对这个方向也蛮喜欢，所以也稍有点动力去学习它。我自认为不是一个特别勤奋，特别聪明的人，但这一年多来，我还是看了不少有关机器学习的书籍和论文，也做了不少相关的笔记，只可惜这些笔记都是手写的。而开通这

2014-03-10 12:17:57 391

Boostable的专栏