Introduction to MATLAB

1
Introduction to MATLAB
Lecture 4 Useful Functions, Debugging, and
Program Organization

• Damon Tomlin
• February 25, 2007

2
Some useful functions

• Initializing variables
• ones, zeros
• Modifying variables
• repmat, reshape
• Statistics
• max, min, mean, std, corrcoef, polyfit, ttest
• Probability
• nchoosek, randperm, rand, randn
• Help
• help, lookfor

3
Program organization -- Subfunctions

• visible only to the main function
• When should I use them?
• Repeated calculations
• Unique to the current program
• Variable management

4
What is debugging?

• A bug is an error in a program
• Bugs come in two flavors
• Those that keep a program from running
• Those that dont
• The first kind is relatively easy because of
  error messages

5
How do we debug a program?

• keyboard (dbquit to return to main workspace)
• dbup dbdown
• profile
• warnings and errors color coded on the right side
  (newer versions)

6
Some useful debugging tools

• disp check on variables
• breakpoints stop a program while it runs
• keyboard the command form of breakpoints
• try catch when an error occurs only sometimes

7
Program organization

• Useful variable names
• Nesting
• Comments

8
Some examples

• b 10
• a 1 1
• c 2 2
• for i 31b
• a(i) sum(a(i-2i-1))
• c(i) sum(c(i-2i-1))
• end

9
Some examples

• NumElements 10
• Start1 1 1 sequence starts with ones
• Start2 2 2 sequence starts with twos
• for i 31NumElements
• Start1(i) sum(Start1(i-2i-1)) element is
  the sum of previous two elements
• Start2(i) sum(Start2(i-2i-1))
• end

10
More on comments

• You can also use commenting to keep parts of your
  program from executing

11
Recursion when a program calls itself

• Searches of unknown depth
• function FoundIt Search(Target)
• search file names for target
• if no target is found
• for each subdirectory
• cd(subdirectory)
• FoundIt Search(Target)
• end
• end
• cd ..

12
Vectorization

• Many loops can be vectorized
• find, repmat, reshape, and element operations
  (. , ./, .) are useful for this
• Matrix1 rand(100) Matrix1 rand(100)
• Matrix2 rand(100) Matrix2 rand(100)
• for i 11size(Matrix1,1) Matrix1(find(Matrix1
  lt Matrix2)) 0
• for j 11size(Matrix1,2)
• if Matrix1(i,j) lt Matrix2(i,j)
• Matrix1(i,j) 0
• end
• end
• end

13
Summary

• MATLAB has built-in functions that help with
  analysis
• Debugging
• Program organization