当前位置:首页 >> 能源/化工 >>

Advanced Statistics with Matlab


Tutorial 5: Advanced Statistics with Matlab

Page 1 of 5

04/22/2004

Tutorial 5? : Advanced Statistics with MATLAB
Daniela Raicu draicu@cs.depaul.edu School of Computer Science, Telecommunications, and Information Systems DePaul University, Chicago, IL 60604
The purpose of this tutorial is to present several advanced statistics techniques using Matlab Statistics toolbox. For this tutorial, we assume that you know the basics of Matlab (covered in Tutorial 1) and the basic statistics functions implemented in Matlab. If you need to refresh your Matlab skills, then a good starting point is to read and solve the exercises from Tutorial #1 (Introduction to Matlab) and Tutorial #3 (Statistics with matlab). The tutorial purpose is to teach you how to use several Matlab built-in functions to calculate advanced statistics for different data sets in different applications; the tutorial is intended for users running a professional version of MATLAB 6.5, Release 13. Topics discussed in this tutorial include: 1. 2. 3. 4. Covariance matrices and Eigenvalues Principal component analysis Canonical Correlation Polynomial fit for a set of points

1. Covariance matrices and Eigenvalues Covariance matrix from raw data “cov.m” ? X = [1 1 2 3;7 8 9 7;1 3 2 1;10 9 11 9;2 2 1 1; 3 4 1 2] X = 1 7 1 10 2 3 ? cov(X) ans = 13.6000 11.6000 15.6000 11.8000
?

1 8 3 9 2 4

2 9 2 11 1 1

3 7 1 9 1 2

11.6000 10.7000 13.6000 9.9000

15.6000 13.6000 19.8667 14.6667

11.8000 9.9000 14.6667 11.3667

Event Sponsor: Visual Computing Area Curriculum Quality of Instruction Council (QIC) grant

Tutorial 5: Advanced Statistics with Matlab

Page 2 of 5

04/22/2004

? cov(X(:,1)) ans = 13.6000 ? cov(X(:,2)) ans = 10.7000 Eigenvalues from raw data “eig.m” ? eig(cov(X)) ans = 0.5731 0.1590 1.4018 %second largest eigenvalue that captures the second largest variance 53.3994 %largest eigenvalue that captures the largest variance in the data 2. Principal Component Analysis components analysis from raw data: “princomp.m”: takes a data matrix X and returns the principal components in PC, the socalled Z-scores in SCORES, the eigenvalues of the covariance matrix of X in LATENT, and Hotelling's T-squared statistic for each data point in TSQUARE. Principal

? [pc, score, latent, tsquare] = princomp(X) pc = 0.4961 0.4316 0.6032 0.4514 -0.3697 -0.6706 0.4179 0.4889 0.6397 -0.4003 -0.5296 0.3874 -0.4561 0.4515 -0.4254 0.6381

score = -4.7824 7.2433 -4.8221 11.2723 -5.3608 -3.5502 2.0736 0.0423 -0.2454 0.0763 -0.3624 -1.5843 0.3947 -0.7265 -1.1806 0.5080 0.3889 0.6155 0.2489 0.2473 -0.1243 -0.2441 -0.6064 0.4786

latent =

Tutorial 5: Advanced Statistics with Matlab

Page 3 of 5

04/22/2004

53.3994 1.4018 0.5731 0.1590

tsquare = 4.1571 2.2893 3.0077 3.2088 3.2088 4.1284

3. Canonical Correlation Canonical correlation analysis for two types of attributes: “Canoncorr.m” [A,B,R,U,V] = CANONCORR(X,Y) computes the sample canonical coefficients for the N-by-D1 and N-by-D2 data matrices X and Y. X and Y must have the same number of observations (rows) but can have different numbers of variables (cols). The jth columns of A and B contain the canonical coefficients, i.e. the linear combination of variables making up the jth canonical variable for X and Y, respectively. Columns of A and B are scaled to make COV(U) and COV(V) (see below) the identity matrix. R containing the sample canonical correlations; the jth element of R is the correlation between the jth columns of U and V (see below), the canonical variables computed as U = (X - repmat(mean(X),N,1))*A and V = (Y - repmat(mean(Y),N,1))*B. Example: load carbig; X = [Displacement Horsepower Weight Acceleration MPG]; nans = sum(isnan(X),2) > 0; [A B r U V] = canoncorr(X(~nans,1:3), X(~nans,4:5)); plot(U(:,1),V(:,1),'.'); xlabel('0.0025*Disp + 0.020*HP - 0.000025*Wgt'); ylabel('-0.17*Accel + -0.092*MPG')

>> A A = 0.0025 0.0202 -0.0000 >> size(X) ans = 0.0048 0.0409 -0.0027

Tutorial 5: Advanced Statistics with Matlab

Page 4 of 5

04/22/2004

406 >> B B = -0.1666 -0.0916 >> r r = 0.8782

5

-0.3637 0.1078

0.6328

4. >> Y =

Polynomial fitting

1 7 1 10 2 3 2

1 8 3 9 2 4 3

Tutorial 5: Advanced Statistics with Matlab

Page 5 of 5

04/22/2004

6

6

? plot(Y(:,2),Y(:,1),'+')

Fit polynomial to data: “polyfit.m”: [P, S] = POLYFIT(X, Y, N) returns the polynomial coefficients P and a structure S for use with POLYVAL to obtain error estimates on predictions. >>[P,S]=polyfit(Y(:,1),Y(:,2),2) P = -0.0309 S = R: [3x3 double] df: 5 normr: 1.9199 Evaluate polynomial: “polyval.m” Y = POLYVAL(P,X), when P is a vector of length N+1 whose elements are the coefficients of a polynomial, is the value of the polynomial evaluated at X. 1.1623 0.6383

Y = P(1)*X^N + P(2)*X^(N-1) + ... + P(N)*X + P(N+1)

If X is a matrix or vector, the polynomial is evaluated at all points in X. See also POLYVALM for evaluation in a matrix sense. [Y, DELTA] = POLYVAL (P,X,S) uses the optional output structure generated by POLYFIT to generate error estimates, Y +/- delta.

>> Y_new=polyval(P,Y(:,1))

Y_new =

1.7697 7.2612 1.7697 9.1731 2.8394 3.8473 2.8394 6.5004 >> plot(Y(:,2),Y(:,1),'*',Y_new(:,1),Y(:,1),'+')



相关文章:
武汉理工大学MATLAB课程设计
The main difference I have introduced in advanced mathematics matlab computer ... histogram statistics and histogram equalization, and join all noises, use a...
Matlab功能应用介绍
Statistics 非线性控制 统计 §1.2 Matlab 基本用法从 Windows 中双击 Matlab 图标,会出现 Matlab 命令窗口(Command Window) ,在一 段提示信息后,出现系统提示符...
Matlab使用方法和简介
Statistics 非线性控制 统计 §1.2 Matlab 基本用法从 Windows 中双击 Matlab 图标,会出现 Matlab 命令窗口(Command Window) ,在一 段提示信息后,出现系统提示符...
Matlab2011b的HMM(隐马尔可夫模型)相关函数介绍
Matlab2011b的HMM(隐马尔可夫模型)相关函数介绍_IT/计算机_专业资料。主要介绍Matlab2011b中的统计工具箱(statistics toolbox)中HMM(隐马尔可夫模型)的相关函数。...
matlab中数据的统计描述和分析
我们将用 Matlab 的统计工具箱(Statistics Toolbox)来实现数据的统计描述和分析。 §1 统计的基本概念 1.1 总体和样本 总体是人们研究对象的全体,又称母体,如...
基于Matlab的语音识别系统的设计
基于Matlab的语音识别系统的设计_自然科学_专业资料。HMM GUI 河南理工大学毕业...target.After statistics,recognition result obviously is made out as the ...
HMM matlab代码实现
HMM matlab 代码实现+分析 收藏 早就想吧 HMM 的代码整理出来,今天找个时间弄... dirichlet) % COMPUTE_ESS_DHMM Compute the Expected Sufficient Statistics for...
LEACH协议代码(MATLAB)
%Values for Hetereogeneity %Percentage of nodes than are advanced m=0....(n+1).yd,'x'); STATISTICS(r+1).DEAD=dead; DEAD(r+1)=dead; DEAD...
leach分簇算法仿真基于matlab
%Values for Hetereogeneity %Percentage of nodes than are advanced m=1;...(dead == n)%节点全部死亡退出循环 break; end STATISTICS(r+1).DEAD=dead...
Matlab 函数分类汇总-R2011b版
Matlab 函数分类汇总-R2011b版_计算机软件及应用_IT/计算机_专业资料。Matlab ...Statistics 10 Distribution Fitting Negative Log-Likelihood Random Number ...
更多相关标签: