|
|
|
發(fā)新文章 |
|
 |
轉(zhuǎn)自prfans：
Sparse, L1-minimization, Compressive Sensing 集中討論帖（第一頁常更新）Sparse大家并不陌生，是個經(jīng)典話題了。而此時 sparse已經(jīng)卷土重來，雖然還是那一鍋湯，但是藥已經(jīng)換了。以 L1-minimization為核心的算法，近幾年飛速進(jìn)展， Compressive Sensing ( Compressive Sampling) 已然成為數(shù)學(xué)領(lǐng)域和信號處理最前沿最熱門的方向。最近一年多這種新形式的算法快速蔓延到模式識別界應(yīng)用， 論文質(zhì)量高、算法效果好、而且算法一般都非常簡單。 而這僅僅是個開始，所以我一直有這個想法專開一貼，供大家一起討論、共同進(jìn)步，今天付諸與行動，希望大家支持。在這個地方（第一個帖），我會陸續(xù)更新提供一些這方面的材料，供大家了解。如果大家提供了有趣的材料，我也盡量加進(jìn)來。當(dāng)然， 此貼重點還是放在理論應(yīng)用和模式識別上。大家踴躍發(fā)言??！ Compressive Sensing資源主頁： Compressive Sensing Resources （最權(quán)威最全面的Compressive Sensing資源主頁，幾乎什么都能找的到）； Compressive Sensing （和上面的差不多）； Compressive Sensing Listing； 馬毅的課程主頁Compressive Sensing Videos； Compressed Sensing Codes （還有 Compressive Sensing Resources 的Software一欄中）； Nuit Blanche； Compressive Sensing: The Big Picture； Terence Tao's What's new; 理論方面的代表人物： David Donoho; Emmanuel Candes; TutorialsEmmanuel Candès, Compressive sampling. (Int. Congress of Mathematics, 3, pp. 1433-1452, Madrid, Spain, 2006) Richard Baraniuk, Compressive sensing. (IEEE Signal Processing Magazine, 24(4), pp. 118-121, July 2007) Emmanuel Candès and Michael Wakin, An introduction to compressive sampling. (IEEE Signal Processing Magazine, 25(2), pp. 21 - 30, March 2008) Justin Romberg, Imaging via compressive sampling. (IEEE Signal Processing Magazine, 25(2), pp. 14 - 20, March 2008) Conferences and SymposiumsShort Course: Sparse Representations and High Dimensional Geometry, May 30 - June 1, 2007 New Directions Short Course: Compressive Sampling and Frontiers in Signal Processing, June 4 - 15, 2007 （ 介紹性的資料和視頻） 理論方面的代表文獻(xiàn)：
Donoho 和 Candes 的文章幾乎都是經(jīng)典模式識別領(lǐng)域的應(yīng)用（包括機(jī)器視覺）： 大家可以去 Compressive Sensing Resources 看 Statistical Signal Processing, Machine Learning, Bayesian Methods, Applications of Compressive Sensing 等欄目 馬毅的一系列論文John Wright, Allen Yang, Arvind Ganesh, Shankar Shastry, and Yi Ma, Robust face recognition via sparse representation. (To appear in IEEE Trans. on Pattern Analysis and Machine Intelligence) , 2008 Allen Yang, John Wright, Yi Ma, and Shankar Sastry, Feature selection in face recognition: A sparse representation perspective. (Preprint, 2007) Kwak, N., Principal Component Analysis Based on L1-Norm Maximization, IEEE Transactions on Pattern Analysis and Machine Intelligence, 2008. Bhusnurmath, Arvind; Taylor, Camillo J., Graph Cuts via $ell_1$ Norm Minimization, IEEE Transactions on Pattern Analysis and Machine Intelligence, 2008. Jianchao Yang, John Wright, Thomas Huang, and Yi Ma, Image Super-Resolution as Sparse Representation of Raw Image Patches, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), June 2008. Arvind Ganesh, Zihan Zhou, and Yi Ma, Separation of A Subspace-Sparse Signal: Algorithms and Conditions, ICASSP 2009.
閱讀文獻(xiàn)的一點心得
1. 先看綜述，后看論著看綜述搞清概念，看論著掌握方法。
2. 早動手在師兄師姐離開之前學(xué)會關(guān)鍵技術(shù)。
3. 多數(shù)文章看摘要，少數(shù)文章看全文掌握了一點查全文的技巧，往往會以搞到全文為樂，以至于沒有時間看文章的內(nèi)容，更不屑于看摘要。真正有用的全文并不多，過分追求全文是浪費，不可走極端。當(dāng)然只看摘要也是不對的。
4. 集中時間看文獻(xiàn)看過總會遺忘?？次墨I(xiàn)的時間越分散，浪費時間越多。集中時間看更容易聯(lián)系起來，形成整體印象。
5. 做好記錄和標(biāo)記復(fù)印或打印的文獻(xiàn)，直接用筆標(biāo)記或批注。pdf 或html 格式的文獻(xiàn)，可以用編輯器標(biāo)亮或改變文字顏色。這是避免時間浪費的又一重要手段。否則等于沒看。
6. 準(zhǔn)備引用的文章要親自看過。轉(zhuǎn)引造成的以訛傳訛不勝枚舉。
7. 注意文章的參考價值?？锏挠绊懸蜃印⑽恼碌谋灰螖?shù)能反映文章的參考價值。但要注意引用這篇文章的其它文章是如何評價這篇文章的：支持還是反對，補(bǔ)充還是糾錯。
8. 交流是最好的老師做實驗遇到困難是家常便飯。你的第一反應(yīng)是什么？反復(fù)嘗試？放棄？看書？這些做法都有道理，但首先應(yīng)該想到的是交流。對有身份的人，私下的請教體現(xiàn)你對他的尊重；對同年資的人，公開的討論可以使大家暢所欲言，而且出言謹(jǐn)慎。千萬不能閉門造車。一個實驗折騰半年，后來別人告訴你那是死路，豈不冤大頭？
9. 最高層次的能力是表達(dá)能力再好的工作最終都要靠別人認(rèn)可。表達(dá)能力，體現(xiàn)為寫和說的能力，是需要長期培養(yǎng)的素質(zhì)。比如發(fā)現(xiàn)一個罕見病例，寫好了發(fā)一篇論著；寫不好只能發(fā)一個病例報道。比如做一個課題，寫好了發(fā)一篇或數(shù)篇論著；寫不好只能發(fā)一個論著摘要或被槍斃。一張圖,一張表,無不是表達(dá)能力的體現(xiàn)。寥寥幾百上千字的標(biāo)書，可以贏得大筆基金；雖然關(guān)系很重要，但寫得太差也不行。有人說，我不學(xué)PCR，不學(xué)spss，只要學(xué)會ppt（powerpoint）就可以了。此話有一點道理，實驗室的boss 們表面上就是靠一串串ppt 行走江湖的。經(jīng)常有研究生因思維敏捷條例清楚而令人肅然起敬。也經(jīng)常有研究生不理解"為什么我做了大部分工作而老板卻讓另一個沒怎么干活的人寫了文章？讓他去大會發(fā)言？"你沒有看到人家有張口就來的本事嗎？
10. 學(xué)好英語，不學(xué)二外。如今不論去日本還是歐洲，學(xué)術(shù)交流早已是英語的天下。你不必為看不懂一篇法語的文章而遺憾，寫那篇文章的人正在為沒學(xué)好英語而犯愁。如果英文尚未精通，暫且不要去學(xué)二外。
文獻(xiàn)管理
1. 下載電子版文獻(xiàn)時（caj，pdf，html），把文章題目粘貼為文件名。注意，文件名不能有特殊符號，要把 \ / : * ? < > | 以及 換行符刪掉。 每次按照同樣的習(xí)慣設(shè)置文件名，可以防止重復(fù)下載。
2. 不同主題存入不同文件夾。文件夾的題目要簡短，如：PD，LTP,PKC，NO。
3. 看過的文獻(xiàn)歸入子文件夾，最起碼要把有用的和沒用的分開。
4. 重要文獻(xiàn)根據(jù)重要程度在文件名前加001,002，003 編號，然后按名稱排列圖標(biāo)，最重要的文獻(xiàn)就排在最前了。
Cover letter:
關(guān)于英文投稿時的cover letter, 以下有三種寫法，各有其特色，但本人認(rèn)為，核心點在于：所做工作的新穎性和關(guān)鍵點，尤其是所做工作解決了什么科學(xué)問題。cover letter 不易過長，關(guān)且注意：這部分是給主編看的！因此，要用最短的文字來說服主編你的文章值得該刊發(fā)表。
附： Case 1Dear Editor, We would like to submit the enclosed manuscript entitled "GDNF Acutely Modulates Neuronal Excitability and A-type Potassium Channels in Midbrain Dopaminergic Neurons", which we wish to be considered for publication in Nature Neuroscience. GDNF has long been thought to be a potent neurotrophic factor for the survival of midbrain dopaminergic neurons, which are degenerated in Parkinson’s disease. In this paper, we report an unexpected, acute effect of GDNF on A-type potassium channels, leading to a potentiation of neuronal excitability, in the dopaminergic neurons in culture as well as in adult brain slices. Further, we show that GDNF regulates the K+ channels through a mechanism that involves activation of MAP kinase. Thus, this study has revealed, for the first time, an acute modulation of ion channels by GDNF. Our findings challenge the classic view of GDNF as a long-term survival factor for midbrain dopaminergic neurons, and suggest that the normal function of GDNF is to regulate neuronal excitability, and consequently dopamine release. These results may also have implications in the treatment of Parkinson’s disease. Due to a direct competition and conflict of interest, we request that Drs. XXX of Harvard Univ., and YY of Yale Univ. not be considered as reviewers. With thanks for your consideration, I am Sincerely yours, case2Dear Editor, We would like to submit the enclosed manuscript entitled "Ca2+-binding protein frequenin mediates GDNF-induced potentiation of Ca2+ channels and transmitter release", which we wish to be considered for publication in Neuron. We believe that two aspects of this manuscript will make it interesting to general readers of Neuron. First, we report that GDNF has a long-term regulatory effect on neurotransmitter release at the neuromuscular synapses. This provides the first physiological evidence for a role of this new family of neurotrophic factors in functional synaptic transmission. Second, we show that the GDNF effect is mediated by enhancing the expression of the Ca2+-binding protein frequenin. Further, GDNF and frequenin facilitate synaptic transmission by enhancing Ca2+ channel activity, leading to an enhancement of Ca2+ influx. Thus, this study has identified, for the first time, a molecular target that mediates the long-term, synaptic action of a neurotrophic factor. Our findings may also have general implications in the cell biology of neurotransmitter release. [0630][投稿寫作]某 雜志給出的標(biāo)準(zhǔn)S ample Cover Letter[the example used is the IJEB] Case 3Sample Cover Letter[the example used is the IJEB] Dear Editor of the [please type in journal title or acronym]: Enclosed is a paper, entitled "Mobile Agents for Network Management." Please accept it as a candidate for publication in the [journal title]. Below are our responses to your submission requirements. 1. Title and the central theme of the article. Paper title: "Mobile Agents for Network Management." This study reviews the concepts of mobile agents and distributed network management system. It proposes a mobile agent-based implementation framework and creates a prototype system to demonstrate the superior performance of a mobile agent-based network over the conventional client-server architecture in a large network environment. 2. Which subject/theme of the Journal the material fits New enabling technologies (if no matching subject/theme, enter 'Subject highly related to [subject of journal] but not listed by [please type in journal title or acronym]) 3. Why the material is important in its field and why the material should be published in [please type in journal title or acronym]? The necessity of having an effective computer network is rapidly growing alongside the implementation of information technology. Finding an appropriate network management system has become increasingly important today's distributed environment. However, the conventional centralized architecture, which routinely requests the status information of local units by the central server, is not sufficient to manage the growing requests. Recently, a new framework that uses mobile agent technology to assist the distributed management has emerged. The mobile agent r educes network traffic, distributes management tasks, and improves operational performance. Given today's bandwidth demand over the Internet, it is important for the [journal title/acronym] readers to understand this technology and its benefits. This study gives a real-life example of how to use mobile agents for distributed network management. It is the first in the literature that reports the analysis of network performance based on an operational prototype of mobile agent-based distributed network. We strongly believe the contribution of this study warrants its publication in the [journal title/acronym]. 4. Names, addresses, and email addresses of four expert referees. Prof. Dr. William Gates Chair Professor of Information Technology 321 Johnson Hall Premier University Lancaster, NY 00012-6666, USA phone: +1-888-888-8888 - fax: +1-888-888-8886 e-mail: wgates@lancaster.edu Expertise: published a related paper ("TCP/IP and OSI: Four Strategies for Interconnection") in CACM, 38(3), pp. 188-198. Relationship: I met Dr. Gate only once at a conference in 1999. I didn't know him personally. Assoc Prof. Dr. John Adams Director of Network Research Center College of Business Australian University 123, Harbor Drive Sydney, Australia 56789 phone: +61-8-8888-8888 - fax: +61-8-8888-8886 e-mail: jadams@au.edu.au Expertise: published a related paper ("Creating Mobile Agents") in IEEE TOSE, 18(8), pp. 88-98. Relationship: None. I have never met Dr. Adams. Assoc Prof. Dr. Chia-Ho Chen Chair of MIS Department College of Management Open University 888, Putong Road Keelung, Taiwan 100 phone: +886-2-8888-8888 - fax: +886-2-8888-8886 e-mail: chchen@ou.edu.tw Expertise: published a related paper ("Network Management for E-Commerce") in IJ Electronic Business, 1(4), pp. 18-28. Relationship: Former professor, dissertation chairman. Mr. Frank Young Partner, ABC Consulting 888, Seashore Highway Won Kok, Kowloon Hong Kong phone: +852-8888-8888 - fax: +852-8888-8886 e-mail: fyoung@abcc.com Expertise: Mr. Young provides consulting services extensively to his clients regarding network management practices. Relationship: I have worked with Mr. Young in several consulting projects in the past three years. Finally, this paper is our original unpublished work and it has not been submitted to any other journal for reviews. Sincerely, Johnny Smith
如何使EndNote顯示所有作者？
菜單里Edit->Output stytles->比如選中Edit "IEEE”，Author lists，選中"List all authors”即可
Remove field codes、Editor如何查找(1) Remove field codes在Endnote X4中,在word2007，Endnote X4菜單，中間Bibilography，在Convert citations and Bibilography的Convert to plain text (2) AI in medicine 這個雜志要求：Add the editors in reference of conference proceedings。怎么找會議的editor呢？
以The 2007 International Conference of Data Mining and Knowledge Engineering為例，怎么查找editor？到該會議主頁，contact us->Editorial board->Editors
總而言之，一個會議的網(wǎng)頁上應(yīng)該包含了你所需要的所有信息，諸如Editors之類。
但是NIPS 2004 沒找到
要求添加Semi-supervised learning using Gaussian fields and harmonic functions.這篇論文的：the editors, the publishing company, and the place of the publishing company in references of conference proceedings.解決方案：到谷歌上搜索www.informatik.uni-trier.de icml 2003 就能找到。
多日想不出此問題的解決方案，經(jīng)過shuling wang老師提醒，終結(jié)解決方案：到web of science上搜，editors和publisher這些信息均能找到，如NIPS上論文learning with local and global consistency和Feature extraction from tumor gene expression profiles using DCT and DFT均能找到
EndNote在LaTeX中的運(yùn)用 本課件幫助科研人員在LaTeX文本編輯環(huán)境下如何利用EndNote軟件編輯參考文獻(xiàn)。
EndNote在LaTeX中的運(yùn)用 （Understand completely） BibTeX Export_Sww.ens.
首次：在Endnote菜單中選擇Edit->Output sytle-> Open style manger-> 選中BibTex export；以后：只要Edit->Output sytle-> BibTex export ，則Endnote中的文獻(xiàn)以Latex的方式呈現(xiàn)
在Endnote中查找那些已經(jīng)在enl中有的文件 在Endnote工具欄中有搜索按鈕，搜索對應(yīng)論文的標(biāo)題即可
4.1 字符串?dāng)?shù)組
4.1.1 字符串入門
【 * 例 4.1.1 -1 】先請讀者實際操作本例，以體會數(shù)值量與字符串的區(qū)別。
clear % 清除所有內(nèi)存變量
a=12345.6789 % 給變量 a 賦數(shù)值標(biāo)量
class(a) % 對變量 a 的類別進(jìn)行判斷
a_s=size(a) % 數(shù)值數(shù)組 a 的“大小”
a =
1.2346e+004
ans =
double
a_s =
1 1
b='S' % 給變量 b 賦字符標(biāo)量（即單個字符）
class(b) % 對變量 b 的類別進(jìn)行判斷
b_s=size(b) % 符號數(shù)組 b 的“大小”
b =
S
ans =
char
b_s =
1 1
whos % 觀察變量 a,b 在內(nèi)存中所占字節(jié)
Name Size Bytes Class
a 1x1 8 double array
a_s 1x2 16 double array
ans 1x4 8 char array
b 1x1 2 char array
b_s 1x2 16 double array
Grand total is 10 elements using 50 bytes
4.1.2 串?dāng)?shù)組的屬性和標(biāo)識
【 * 例 4.1.2 -1 】本例演示：串的基本屬性、標(biāo)識和簡單操作。
（1）創(chuàng)建串?dāng)?shù)組
a='This is an example.'
a =
This is an example.
（2）串?dāng)?shù)組 a 的大小
size(a)
ans =
1 19
（3）串?dāng)?shù)組的元素標(biāo)識
a14=a(1:4) % 提出一個子字符串
ra=a(end:-1:1) % 字符串的倒排
a14 =
This
ra =
.elpmaxe na si sihT
（4）串?dāng)?shù)組的 ASCII 碼
ascii_a=double(a) % 產(chǎn)生 ASCII 碼
ascii_a =
Columns 1 through 12
84 104 105 115 32 105 115 32 97 110 32 101
Columns 13 through 19
120 97 109 112 108 101 46
char(ascii_a) % 把 ASCII 碼變回字符串
ans =
This is an example.
（5）對字符串 ASCII 碼數(shù)組的操作
% 使字符串中字母全部大寫
w=find(a>='a'&a<='z'); % 找出串?dāng)?shù)組 a 中，小寫字母的元素位置。
ascii_a(w)=ascii_a(w)-32; % 大小寫字母 ASCII 值差 32. 用數(shù)值加法改變部分碼值。
char(ascii_a) % 把新的 ASCII 碼翻成字符
ans =
THIS IS AN EXAMPLE.
（6）中文字符串?dāng)?shù)組
A=' 這是一個算例。 '; % 創(chuàng)建中文字符串
A_s=size(A) % 串?dāng)?shù)組的大小
A56=A([5 6]) % 取串的子數(shù)組
ASCII_A=double(A) % 獲取 ASCII 碼
A_s =
1 7
A56 =
算例
ASCII_A =
Columns 1 through 6
54754 51911 53947 47350 52195 49405
Column 7
41379
char(ASCII_A) % 把 ASCII 碼翻譯成字符
ans =
這是一個算例。
（7）創(chuàng)建帶單引號的字符串
b='Example '' 4.1.2 -1'''
b =
Example ' 4.1.2 -1'
（8）由小串構(gòu)成長串
ab=[a(1:7),' ',b,' .'] % 這里第 2 個輸入為空格串
ab =
This is Example ' 4.1.2 -1' .
4.1.3 復(fù)雜串?dāng)?shù)組的創(chuàng)建
4.1.3.1 多行串?dāng)?shù)組的直接創(chuàng)建
【 * 例 4.1.3 .1-1 】多行串?dāng)?shù)組的直接輸入示例。
clear
S=['This string array '
'has multiple rows.']
S =
This string array
has multiple rows.
size(S)
ans =
18
4.1.3.2 利用串操作函數(shù)創(chuàng)建多行串?dāng)?shù)組
【 * 例 4.1.3 .2-1 】演示：用專門函數(shù) char , str2mat , strvcat 創(chuàng)建多行串?dāng)?shù)組示例。
S1=char('This string array','has two rows.')
S1 =
This string array
has two rows.
S2=str2mat(' 這 ',' 字符 ',' 串?dāng)?shù)組 ',' 由 4 行組成 ')
S2 =
這
字符
串?dāng)?shù)組
由4 行組成
S3=strvcat(' 這 ',' 字符 ',' 串?dāng)?shù)組 ',' ',' 由 4 行組成 ')% “空串”會產(chǎn)生一個空格行
S3 =
這
字符
串?dāng)?shù)組
由 4 行組成
size(S3)
ans =
5 5
【 * 例 4.1.3 .2-1 】的補(bǔ)充
(1) 創(chuàng)建一個二維字符數(shù)組animal
>> Animal=[‘dog’;’monkey’];
??? Error using ==> vertcat
CAT arguments dimensions are not consistent.
>> Animal=['dog ';'monkey']; %創(chuàng)建成功
說明：創(chuàng)建二維字符數(shù)組時，字符數(shù)組要求每行字符含有相同的列。當(dāng)多行字符串具有不同長度時，為了避免出現(xiàn)錯誤，用戶需要在較短的字符串中添加空格，以便保證較短字符串與最長字符串等長。
(2) 用char函數(shù)創(chuàng)建字符數(shù)組，該方法不需要所有字符串等長
>> Animal = char(‘dog’,’monkey’);
4.1.3.3 轉(zhuǎn)換函數(shù)產(chǎn)生數(shù)碼字符串
【 * 例 4.1.3 .3-1 】最常用的數(shù)組 / 字符串轉(zhuǎn)換函數(shù) int2str , num2str , mat2str 示例。
（1） int2str 把整數(shù)數(shù)組轉(zhuǎn)換成串?dāng)?shù)組（非整數(shù)將被四舍五入園整后再轉(zhuǎn)換）
A=eye(2,4); % 生成一個  數(shù)值數(shù)組
A_str1=int2str(A) % 轉(zhuǎn)換成  串?dāng)?shù)組。請讀者自己用 size 檢驗。
A_str1 =
1 0 0 0
0 1 0 0
（2） num2str 把非整數(shù)數(shù)組轉(zhuǎn)換為串?dāng)?shù)組（常用于圖形中，數(shù)據(jù)點的標(biāo)識）
rand('state',0)
B=rand(2,4); % 生成數(shù)值矩陣
B3=num2str(B,3) % 保持 3 位有效數(shù)字，轉(zhuǎn)換為串
B3 =
0.95 0.607 0.891 0.456
0.231 0.486 0.762 0.0185
（3） mat2str 把數(shù)值數(shù)組轉(zhuǎn)換成輸入形態(tài)的串?dāng)?shù)組（常與 eval 指令配用）
B_str=mat2str(B,4) % 保持 4 位有效數(shù)字，轉(zhuǎn)換為“數(shù)組輸入形式”串
B_str =
[0.9501 0.6068 0.8913 0.4565;0.2311 0.486 0.7621 0.0185]
Expression=['exp(-',B_str,')']; % 相當(dāng)于指令窗寫一個表達(dá)式 exp(-B_str)
eval(Expression) % 把 exp(-B_str) 送去執(zhí)行
ans =
0.3867 0.5451 0.4101 0.6335
0.7937 0.6151 0.4667 0.9817
【 * 例 4.1.3 .3-2 】綜合例題：在 MATLAB 計算生成的圖形上標(biāo)出圖名和最大值點坐標(biāo)。
clear % 清除內(nèi)存中的所有變量
a=2; % 設(shè)置衰減系數(shù)
w=3; % 設(shè)置振蕩頻率
t=0:0.01:10; % 取自變量采樣數(shù)組
y=exp(-a*t).*sin(w*t); % 計算函數(shù)值，產(chǎn)生函數(shù)數(shù)組
[y_max,i_max]=max(y); % 找最大值元素位置
t_text=['t=',num2str(t(i_max))]; % 生成最大值點的橫坐標(biāo)字符串 <7>
y_text=['y=',num2str(y_max)]; % 生成最大值點的縱坐標(biāo)字符串 <8>
max_text=char('maximum',t_text,y_text);% 生成標(biāo)志最大值點的字符串 <9>
% 生成標(biāo)志圖名用的字符串
tit=['y=exp(-',num2str(a),'t)*sin(',num2str(w),'t)']; %<11>
plot(t,zeros(size(t)),'k') % 畫縱坐標(biāo)為 0 的基準(zhǔn)線
hold on % 保持繪制的線不被清除
plot(t,y,'b') % 用蘭色畫 y(t) 曲線
plot(t(i_max),y_max,'r.','MarkerSize',20) % 用大紅點標(biāo)最大值點
text(t(i_max)+0.3,y_max+0.05,max_text) % 在圖上書寫最大值點的數(shù)據(jù)值 <16>
title(tit),xlabel('t'),ylabel('y'),hold off% 書寫圖名、橫坐標(biāo)名、縱坐標(biāo)名
圖 4.1.3 .3-1 字符串運(yùn)用示意圖
4.1.3.4 利用元胞數(shù)組創(chuàng)建復(fù)雜字符串
【 * 例 4.1.3 .4-1 】元胞數(shù)組在存放和操作字符串上的應(yīng)用。
a='MATLAB 5 ';b='introduces new data types:'; % 創(chuàng)建單行字符串 a,b
c1=' ◆ Multidimensional array';c2=' ◆ User-definable data structure';
c3=' ◆ Cell arrays';c4=' ◆ Character array';
c=char(c1,c2,c3,c4); % 創(chuàng)建多行字符串 c
C={a;b;c}; % 利用元胞數(shù)組存放長短不同的字符串 <5>
disp([C{1:2}]) % 顯示前兩個元胞中的字符內(nèi)容 <6>
disp(' ') % 顯示一行空白
disp(C{3}) % 顯示第 3 個元胞中的字符內(nèi)容 <8>
MATLAB 5 introduces new data types:
◆ Multidimensional array
◆ User-definable data structure
◆ Cell arrays
◆ Character array
4.1.4 串轉(zhuǎn)換函數(shù)
【 * 例 4.1.4 -1 】 fprintf, sprintf, sscanf 的用法示例。
rand('state',0);a=rand(2,2); % 產(chǎn)生  隨機(jī)陣
s1=num2str(a) % 把數(shù)值數(shù)組轉(zhuǎn)換為串?dāng)?shù)組
s_s=sprintf('%.10e\n',a) %10 數(shù)位科學(xué)記述串 , 每寫一個元素就換行。
s1 =
0.95013 0.60684
0.23114 0.48598
s_s =
9.5012928515e-001
2.3113851357e-001
6.0684258354e-001
4.8598246871e-001
fprintf('% .5g \\',a) % 以 5 位數(shù)位最短形式顯示。不能賦值用
0.95013\0.23114\0.60684\0.48598\
s_sscan=sscanf(s_s,'%f',[3,2])% 浮點格式把串轉(zhuǎn)換成成  數(shù)值數(shù)組。
s_sscan =
0.9501 0.4860
0.2311 0
0.6068 0
關(guān)于四分位數(shù)(Quartile)在wikipedia維基百科 上搜索英語版或者中文版都有很清晰的解釋
在 wikipedia維基百科 上搜索 Box Plot ：
箱形圖（Box-plot）又稱為盒須圖、盒式圖或箱線圖，是一種用作顯示一組數(shù)據(jù)分散情況資料的統(tǒng)計圖。因型狀如箱子而得名。在各種領(lǐng)域也經(jīng)常被使用，常見于品質(zhì)管理。不過作法相對較較繁瑣。
箱形圖于1977年由美國著名統(tǒng)計學(xué)家 John Tukey發(fā)明。它能顯示出一組數(shù)據(jù)的最大值、最少值、中位數(shù)、下四分位數(shù)及上四分位數(shù)。
以下是箱形圖的具體例子：
+-----+-+
* o |-------| + | |---|
+-----+-+
+---+---+---+---+---+---+---+---+---+---+ 數(shù)線
0 1 2 3 4 5 6 7 8 9 10
這組數(shù)據(jù)顯示出：
- 最小值(min)=5。
- 下四分位數(shù)(Q1)=7。
- 中位數(shù)(Med)=8.5。
- 上四分位數(shù)(Q3)=9。
- 最大值(max)=10。
- 平均值=8。
- 四分位間距(interquartile range)=Q3 − Q1=2
http://www.physics.csbsju.edu/stats/box2.html
The box plot (a.k.a. box and whisker diagram) is a standardized way of displaying the distribution of data based on the five number summary: minimum, first quartile, median, third quartile, and maximum. In the simplest box plot the central rectangle spans the first quartile to the third quartile (the interquartile range or IQR). A segment inside the rectangle shows the median and "whiskers" above and below the box show the locations of the minimum and maximum.
This simplest possible box plot displays the full range of variation (from min to max), the likely range of variation (the IQR), and a typical value (the median). Not uncommonly real datasets will display surprisingly high maximums or surprisingly low minimums called outliers. John Tukey has provided a precise definition for two types of outliers:
- Outliers are either 3×IQR or more above the third quartile or 3×IQR or more below the first quartile.
- Suspected outliers are are slightly more central versions of outliers: either 1.5×IQR or more above the third quartile or 1.5×IQR or more below the first quartile.
If either type of outlier is present the whisker on the appropriate side is taken to 1.5×IQR from the quartile (the "inner fence") rather than the max or min, and individual outlying data points are displayed as unfilled circles (for suspected outliers) or filled circles (for outliers). (The "outer fence" is 3×IQR from the quartile.)
If the data happens to be normally distributed,
IQR = 1.35 σ
where σ is the population standard deviation.
Suspected outliers are not uncommon in large normally distributed datasets (say more than 100 data-points). Outliers are expected in normally distributed datasets with more than about 10,000 data-points. Here is an example of 1000 normally distributed data displayed as a box plot:
Note that outliers are not necessarily "bad" data-points; indeed they may well be the most important, most information rich, part of the dataset. Under no circumstances should they be automatically removed from the dataset. Outliers may deserve special consideration: they may be the key to the phenomenon under study or the result of human blunders.
Example A
Consider two datasets:
A1={0.22, -0.87, -2.39, -1.79, 0.37, -1.54, 1.28, -0.31, -0.74, 1.72, 0.38, -0.17, -0.62, -1.10, 0.30, 0.15, 2.30, 0.19, -0.50, -0.09}
A2={-5.13, -2.19, -2.43, -3.83, 0.50, -3.25, 4.32, 1.63, 5.18, -0.43, 7.11, 4.87, -3.10, -5.81, 3.76, 6.31, 2.58, 0.07, 5.76, 3.50}
Notice that both datasets are approximately balanced around zero; evidently the mean in both cases is "near" zero. However there is substantially more variation in A2 which ranges approximately from -6 to 6 whereas A1 ranges approximately from -2½ to 2½.
Below find box plots and the more traditional error bar plots (with 1-σ bars). Notice the difference in scales: since the box plot is displaying the full range of variation, the y-range must be expanded.
Example B
B1={1.26, 0.34, 0.70, 1.75, 50.57, 1.55, 0.08, 0.42, 0.50, 3.20, 0.15, 0.49, 0.95, 0.24, 1.37, 0.17, 6.98, 0.10, 0.94, 0.38}
B2= {2.37, 2.16, 14.82, 1.73, 41.04, 0.23, 1.32, 2.91, 39.41, 0.11, 27.44, 4.51, 0.51, 4.50, 0.18, 14.68, 4.66, 1.30, 2.06, 1.19}
Notice that the datasets span much the same range of values (from about .1 to about 50) and that all the values are positive. Most of the B1 values are less than one whereas most of the B2 values are more than one. We can use a log scale to better display this large range of values:
On the other hand, a straightforward plot of the sample means and population standard deviations, suggests negative values (which prevents use of a log-scale) and broad overlap between the two distributions. (A t-test would suggest B1 and B2 are not significantly different.)
Example C
One case of particular concern --where a box plot can be deceptive-- is when the data are distributed into "two lumps" rather than the "one lump" cases we've considered so far.
A "bee swarm" plot shows that in this dataset there are lots of data near 10 and 15 but relatively few in between. See that a box plot would not give you any evidence of this.
Matlab中有關(guān)boxplot(X)命令的解釋：
boxplot(X) produces a box and whisker plot for each column of the matrix X. The box has lines at the lower quartile, median, and upper quartile values. Whiskers extend from each end of the box to the adjacent values in the data—by default, the most extreme values within 1.5 times the interquartile range from the ends of the box. Outliers are data with values beyond the ends of the whiskers. Outliers are displayed with a red + sign.
格式 boxplot(X) %產(chǎn)生矩陣X的每一列的盒圖和“須”圖，“須”是從盒的尾部延伸出來，并表示盒外數(shù)據(jù)長度的線，如果“須”的外面沒有數(shù)據(jù)，則在“須”的底部有一個點。
boxplot(X,notch) %當(dāng)notch=1時，產(chǎn)生一凹盒圖，notch=0時產(chǎn)生一矩箱圖。
boxplot(X,notch,'sym') %sym表示圖形符號，默認(rèn)值為“+”。
boxplot(X,notch,'sym',vert) %當(dāng)vert=0時，生成水平盒圖，vert=1時，生成豎直盒圖（默認(rèn)值vert=1）。
boxplot(X,notch,'sym',vert,whis) %whis定義“須”圖的長度，默認(rèn)值為1.5，若whis=0則boxplot函數(shù)通過繪制sym符號圖來顯示盒外的所有數(shù)據(jù)值。
Examples 1
The following commands create a box plot of car mileage grouped by country.
load carsmall
boxplot(MPG,Origin)
Examples 2
The following example produces notched box plots for two groups of sample data.
x1 = normrnd(5,1,100,1);
x2 = normrnd(6,1,100,1);
boxplot([x1,x2],'notch','on')
Examples 3
x1 = normrnd(5,1,100,1);
x2 = normrnd(6,1,100,1);
boxplot([x1,x2])
The difference between the medians of the two groups is approximately 1.Since the notches in the boxplot do not overlap, you can conclude, with 95% confidence, that the true medians do differ.
Examples 4
The following figure shows the boxplot for same data with the length of the whiskers specified as 1.0 times the interquartile range. Points beyond the whiskers are displayed using +.
x1 = normrnd(5,1,100,1); x2 = normrnd(6,1,100,1); boxplot([x1,x2],'notch','on','whisker',1)
版權(quán)聲明：轉(zhuǎn)載時請以超鏈接形式標(biāo)明文章原始出處和作者信息及本聲明
http://netessays.blogbus.com/logs/30974661.html
Endnote里的文獻(xiàn)類型有很多，有journal article、conference paper等。其中，關(guān)于會議的文獻(xiàn)類型有conference proceeding和conference paper兩個，這兩
個有什么區(qū)別呢？
Endnote官方網(wǎng)站上有如下描述：
The Conference Proceedings reference type is best used for unpublished
proceedings. Articles that are published as part of the comprehensive
conference proceedings should be entered as Conference Paper references.
也就是說，對于一般已經(jīng)出版的proceeding，應(yīng)該歸結(jié)到Conference Paper里面，只有沒有出版的Proceeding才放到conference proceeding。
參考文獻(xiàn)樣式： （1） LNAI，LNCS是雜志，會議，會議錄？怎么寫參考文獻(xiàn)格式，按照什么格式Hongqiang Wang 師兄主頁書中的章還是Chunhou Zheng 師兄主頁第六和第八個參考文獻(xiàn)。如Feature extraction from tumor gene expression profiles using DCT and DFT到web of science上可以查到是個會議論文，會議名字也能查到。在Book Series: LECTURE NOTES IN ARTIFICIAL INTELLIGENCE中，Shuling Wang老師講應(yīng)該按照會議論文格式
參考文獻(xiàn)樣式：
比較好的方式：作者，“文章名，”雜志或會議名，卷號（會議地點），期號（會議時間），頁碼，年份。（參見March 10，2010郵件）
(2)Riemannian manifold learning: 查老師這篇文章的第48篇引用文獻(xiàn)是來自維基百科，51和52是引用的數(shù)據(jù)，值得學(xué)習(xí)這種引用方式！
    不多說了,在 金士頓DDRII 667或者DDRII 800 內(nèi)存的真假辨別的這篇文章中已經(jīng)說的很明白了,由于有的玩家朋友希望能拿個假貨進(jìn)行對比,所以這次發(fā)個全套照片供大家查看辨別.
方法：到google上搜索關(guān)鍵詞：CVPR 2009 papers on the web NIPS ( http://books.nips.cc/) ICML ( http://www.cs.mcgill.ca/~icml2009/abstracts.html.). AAAI10: http://www.aaai.org/ocs/index.php/AAAI/AAAI10/schedConf/presentations; AAAI 12: http://www.aaai.org/ocs/index.php/AAAI/AAAI12/schedConf/presentations；AAAI短文在電腦"paper\aa_other\AAAI短文\" ICCV 2019 https://mp.weixin.qq.com/s/-l9Wyh945k3XNeHw-ApjzA
http://openaccess.thecvf.com/ICCV2019.py
|