%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % Matlab Demos % (adapted from http://www.stanford.edu/class/cs223b/matlabIntro.html) % % Stefan Roth , 09/10/2003 % Alex Balan , 09/14/2004 % Leonid Sigal, 09/14/2005 % Greg Shakhnarovich 09/04/2006 % % Last modified: 09/04/2006 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % The symbol "%" is used to indicate a comment (for the remainder of % the line). % Within Matlab interpreter window, you can evaluate a portion of code by % selecting it and pressing F9. % help is useful; even if you don't immediately see the information % you are looking for, the "SEE ALSO" pointers often lead to it. help help help general % this is A = [1, 2; % note: once the matrix input has started, Matlab allows you % to break the line, and correctly inteprets the rest of the input 3, 4]; % A semicolon at the end of a statement means that Matlab will not % display the result of the evaluated statement. If the ";" is omitted % then Matlab will display the result. This is also useful for % printing the value of variables, e.g. A % Matlab's command line is a little like a standard shell: % - Use the up arrow to recall commands without retyping them (and % down arrow to go forward in the command history). % - In linux, C-a moves to beginning of line (C-e for end), C-f moves % forward a character and C-b moves back (equivalent to the left and % right arrow keys), C-d deletes a character, C-k deletes the rest of % the line to the right of the cursor, C-p goes back through the % command history and C-n goes forward (equivalent to up and down % arrows), Tab tries to complete a command. % You can execute a linux command in matlab by prepending "!" to it !ls -l !touch testing !rm testing % Simple debugging: % If the command "dbstop if error" is issued before running a script % or a function that causes a run-time error, the execution will stop % at the point where the error occurred. Very useful for tracking down % errors. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % The basic types in Matlab are scalars (usually double-precision % floating point), vectors, and matrices: N = 5 % A scalar A = [1 2; 3 4] % Creates a 2x2 matrix B = [1 2; 3 4] % The simplest way to create a matrix is % to list its entries in square brackets. % The ";" symbol separates rows; % the (optional) "," separates columns. v = [1 0 0] % A row vector v = [1; 2; 3] % A column vector v = v' % Transpose a vector (row to column or % column to row) v = 1:.5:3 % A vector filled in a specified range: v = pi*[-4:4]/4 + 2 % [start:stepsize:end] % default stepsize is 1 % (brackets are optional) v = [] % Empty vector isempty(v) isempty([1 2]) b = zeros(3, 2) % Matrix of zeros c = ones(2, 3) % Matrix of ones d = eye(4) % Identity matrix d2 = diag([1:4]/4) % Diagonal matrix e = rand(64, 128); % Random matrix (uniform) between [0,1] f = randn(64, 128); % Random matrix (normal) N(0,1) g = zeros(128); % Matrix of zeros (128x128) %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % Visualization: plotting etc. x = [0 2 1 3 4]; % Basic plotting plot(x); % Plot x versus its index values starting from 1 pause % Wait for key press figure; % Open new figure. % by default, Matlab will plot using current figure. % figures are identified by handles -- integers starting from 1. % New figure is created with the lowest available handle. x = pi*[-1:1/24:1] % Vector of arguments for functions to be ploted plot(x, sin(x)); xlabel('radians'); % Assign label for x-axis ylabel('sin value'); % Assign label for y-axis title('dummy title'); % Assign plot title figure(1); % Use the figure with handle 1 for % subsequent plotting commands subplot(1, 2, 1); % Multiple functions in separate graphs plot(x, sin(x)); % (see "help subplot") axis square; % Make visible area square subplot(1, 2, 2); plot(x, 2*cos(x)); axis square; figure(2); % Use the figure with handle 2 plot(x, sin(x)); hold on; % Multiple functions in single graph plot(x, 2*cos(x), '--'); % '--' chooses different line pattern legend('sin', 'cos'); % Assigns names to each plot hold off; % Stop putting multiple figures in current % graph pause close all %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % (C) Indexing vectors and matrices. % Warning: Indices always start at 1 and *NOT* at 0! v = [1 2 3] v(3) % Access a vector element v(5) % this causes an error (out of range) % however, it is OK to add an element: v(5)=5 % note how the 4th element got the default value of zero m = [1 2 3 4; 5 7 8 8; 9 10 11 12; 13 14 15 16] m(3, 2) % Access a matrix element % matrix(ROW #, COLUMN #) % a range of indices can be specified. Note the row transposition effect % due to the negative step -1 in the range expression: m(3:-1:1, [1 3]) % a matrix can be visualized using image or imagesc commands figure; imagesc(m) % colors correspond to values according to colormap (help colormap) % the color "legend" can be displayed explicitly: colorbar % more matix stuff: m = rand(128,256); size(m) % Returns the size of a matrix size(m, 1) % Number of rows size(m, 2) % Number of columns length(m) % max(size(m)) m1 = zeros(size(m)); % Create a new matrix with the size of m, % fill it with zeros % ---- Simple operations on vectors and matrices % ---- Reshaping and assembling matrices: a = [1 2; 3 4; 5 6] % A 3x2 matrix b = a(:) % Make 6x1 column vector by stacking % up columns of a sum(a(:)) % Useful: sum of all elements c = reshape(a, [2,3]) % Make 2 x 3 matrix out of a 3 x 2 matrix % elements are arranged column-major % note: reshape will throw error if sizes % don't add up % Matrix concatenation d = [c -c] e = repmat(a, 2, 2) % Create a 3x2 replication of the image p %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % (B) Vector Operations % Built-in Matlab functions that operate on vectors a = [1 4 6 3] % A row vector sum(a) % Sum of vector elements mean(a) % Mean of vector elements var(a) % Variance of elements std(a) % Standard deviation max(a) % Maximum min(a) % Minimum % If a matrix is given, then these functions will operate on each column % of the matrix and return a row vector as result a = [1 2 3; 4 5 6] % A matrix mean(a) % Mean of each column max(a) % Max of each column max(max(a)) % Obtaining the max of a matrix mean(a, 2) % Mean of each row (second argument specifies % dimension along which operation is taken) sum(a,2) % sum of each row sum(a) % the default dimension is always 1 (i.e., % sum of each column in this case). % Note: the result of such an operation is % normally a vector! % ' performs matrix transposition [1 2 3] * [4 5 6]' % 1x3 row vector times a 3x1 column vector % results in a scalar. Known as dot product % or inner product. Note the absence of "." [1 2 3]' * [4 5 6] % 3x1 column vector times a 1x3 row vector % results in a 3x3 matrix. Known as outer % product. Note the absence of "." % Other elementary functions: % >> abs(x) % absolute value of x % >> exp(x) % e to the x-th power % >> fix(x) % rounds x to integer towards 0 % >> log10(x) % common logarithm of x to the base 10 % >> rem(x,y) % remainder of x/y % >> sqrt(x) % square root of x % >> sin(x) % sine of x; x in radians % >> acoth(x) % inversion hyperbolic cotangent of x % Other element-wise arithmetic operations include e.g. : % floor, ceil, ... help elfun % get a list of all available elementary functions % ---- element-wise versus matrix operations: a = [1 2 3; 4 5 6] b = [10 20 30; 40 50 60] % "x . b" means perform operation on each pair of matching (by % indices) elements of a and b. This will cause error if the sizes are % not equal. % compare this command a.*b % to this (illegal) command a*b b./5 % if one of the arguments is a scalar, it is used for each element: a.^2 2.^a %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % Other data structures: % "Structures" are multidimensional Matlab arrays with elements accessed % by textual field designators. For example, A.name = 'Rudi'; A.score = [ 2 3 4]; A.fancy = rand; A % Like everything else in Matlab, structures are arrays, % so you can insert additional elements. A(2) = A(1); A(2).name = 'Fredi'; A(1) A(2) %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % "Cell Arrays" % Cell arrays in Matlab are multidimensional arrays whose elements can % be anything. Cell arrays are created by enclosing a % miscellaneous collection of things in curly braces, fg: S = 'Hello World!' R = rand(2,3,2) C = {R -1 S} % To retrieve the contents of one of the cells, use subscripts in % curly braces. C{3} % Like numeric arrays also cell arrays can be mulitdimensional CM{2,2} = S; CM{1,1} = 'hallo'; CM %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % (4) Control statements & vectorization % Syntax of control flow statements: % % for VARIABLE = EXPR % STATEMENT % ... % STATEMENT % end % % EXPR is a row vector here, e.g. 1:10 or -1:0.5:1 or [1 4 7] % - it should never be a column vector % % % while EXPRESSION % STATEMENTS % end % % if EXPRESSION % STATEMENTS % elseif EXPRESSION % STATEMENTS % else % STATEMENTS % end % % (elseif and else clauses are optional, the "end" is required) % % EXPRESSIONs are usually made of relational clauses, e.g. a < b % The operators are <, >, <=, >=, ==, ~= (almost like in C(++)) % % Element-wise Logical operators : % and: & % or: | % not: ~ % exclusive or: xor % For short-circuit: &&, || % Warning: % Loops run very slowly in Matlab, because of interpretation overhead. % This has gotten somewhat better in recent versions, but you should % nevertheless try to avoid them by "vectorizing" the computation, % i.e. by rewriting the code in form of matrix operations. This is % illustrated in some examples below. R = [4 5] > [3 7] any(R) % True if any element of a vector is nonzero all(R) % True if all elements of a vector are nonzero % Examples: for i=1:2:7 % Loop from 1 to 7 in steps of 2 i % Print i end for i=[5 13 -1] % Loop over given vector if (i > 10) % Sample if statement disp('Larger than 10') % Print given string elseif i < 0 % Parentheses are optional disp('Negative value') else disp('Something else') end end % Here is another example: given an mxn matrix A and a 1xn % vector v, we want to subtract v from every row of A. m = 50000; n = 10; A = ones(m, n); v = 2 * rand(1, n); % % Implementation using loops: tic % start stopwatch for i=1:m A(i,:) = A(i,:) - v; end toc % stop stopwatch % We can compute the same thing using only matrix operations tic % start stopwatch A = ones(m, n) - repmat(v, m, 1); % This version of the code runs % much faster!!! toc % stop stopwatch % We can vectorize the computation even when loops contain % conditional statements. % % Example: given an mxn matrix A, create a matrix B of the same size % containing all zeros, and then copy into B the elements of A that % are greater than zero. % Implementation using loops: tic % start stopwatch B = zeros(m,n); for i=1:m for j=1:n if A(i,j)>0 B(i,j) = A(i,j); end end end toc % stop stopwatch % All this can be computed w/o any loop! tic % start stopwatch B = zeros(m,n); ind = find(A > 0); % Find indices of positive elements of A % (see "help find" for more info) B(ind) = A(ind); % Copies into B only the elements of A % that are > 0 toc % stop stopwatch % The indeces "find" returns can be converted back to row-column index % using ind2sub. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % fprintf : Can be used to print output to the screen or to a file. % Print to the screen x=pi; fprintf('\nThis is a test, x=%g \n',x) % This command prints the number x in % either exponential or fixed point formant % (whichever is shorter) and then starts a new % line. disp(x) disp(['X = ' num2str(x)]) % Display by string concatenation % Print to a file called 'output.txt': fid=fopen('output.txt') fprintf(fid,'hello %g', x) % Get input from the keyboard, e.g., a = input('Left Endpoint a = ?') %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %(6) Creating scripts or functions using m-files: % % Matlab scripts are files with ".m" extension containing Matlab % commands. Variables in a script file are global and will change the % value of variables of the same name in the environment of the current % Matlab session. A script with name "script1.m" can be invoked by % typing "script1" in the command window. % Functions are also m-files. The first line in a function file must be % of this form: % function [outarg_1, ..., outarg_m] = myfunction(inarg_1, ..., inarg_n) % % The function name should be the same as that of the file % (i.e. function "myfunction" should be saved in file "myfunction.m"). % Have a look at myfunction.m and myotherfunction.m for examples. % % Functions are executed using local workspaces: there is no risk of % conflicts with the variables in the main workspace. At the end of a % function execution only the output arguments will be visible in the % main workspace. a = [1 2 3 4]; % Global variable a b = myfunction(a .^ 2) % Call myfunction which has local % variable a a % Global variable a is unchanged % NOTE: % When writing a long Matlab statement that becomes too long for a % single line use "..." at the end of the line to continue on the next % line. E.g. [c, d] = ... myotherfunction(a, b) % Call myotherfunction with two return % values % %%%In "myfunction.m" %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % function y = myfunction(x) % % Function of one argument with one return value % % a = [1 1 1 1]; % Have a global variable of the same name % y = a + x; % %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % %%%In "myotherfunction.m" %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % function [y, z] = myotherfunction(a, b) % % Function of two arguments with two return values % % y = a + b; % z = a - b; % %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %(5) Saving your work who % List variables in workspace whos % List variables w/ info about size, type, etc. save myfile % Saves all workspace variables into % file myfile.mat save myfile a b % Saves only variables a and b clear a b % Removes variables a and b from the % workspace clear % Clears the entire workspace load myfile a % Loads variable a load myfile % Loads variable(s) from myfile.mat %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%