MATLAB and R Programming (ME-705) - Unit 2 Short Notes

UNIT 2: PROGRAMMING FUNDAMENTALS & DATA MANIPULATION

2.1 Core Programming Constructs & Control Flow

Conditional Statements

• MATLAB: if, elseif, else, switch (with case).

• R: if, else if, else; switch() for multiple values; ifelse() for vectorized conditionals.

• [!TIP] ifelse() in R is vectorized—operates on entire vectors without loops. MATLAB’s if is scalar; use logical indexing for vectorized conditionals.

Loops

• for loops: iterate over indices (for i=1:n) or values (for val in array).

• while loops: repeat while condition true.

• Loop control: break (exit loop), continue (skip to next iteration).

• [!CAUTION] Nested loops increase time complexity to $$\displaystyle O(n^2) $$; avoid for large datasets.

Logical & Relational Operators

Vectorization

• Definition: Performing operations on entire arrays/matrices directly, avoiding explicit loops.

• Benefit: Leverages optimized BLAS/LAPACK libraries; significantly faster (often $O(n)$ vs. $O(n)$ with lower constant factor for loops).

• Example (MATLAB): C = A + B; vs. for i=1:length(A), C(i)=A(i)+B(i); end.

• Example (R): x * 2 vs. for(i in 1:length(x)) x[i] <- x[i]*2.

• [!TIP] Exam Focus: Always prefer vectorized operations in MATLAB; in R, use apply() family or vectorized functions like ifelse().

2.2 Functions & Modular Programming

Defining Functions

• MATLAB:

  • Function file: function [out1, out2] = name(in1, in2) ... end.

  • Anonymous function: f = @(x) x.^2;.

• R:

  • name <- function(arg1, arg2) { ... return(out) }.

  • No separate file required; can be defined in scripts.

Function Components

• Input arguments: Passed by value (copy-on-write in MATLAB; R uses pass-by-promise).

• Output/return: MATLAB: assign to function name or use return(); R: last expression or return().

• Scope: Variables inside function are local by default. Use global (MATLAB) or <<- (R) for global (avoid if possible).

• [!CAUTION] Modifying global variables leads to hard-to-debug code; prefer returning values.

Calling Functions

• Positional arguments: Order matters.

• Named arguments (R only): fun(arg1=val1, arg2=val2)—improves readability.

Built-in Functions

• Mathematical: sin, cos, exp, log (both).

• Statistical: mean, median, std, var (both); summary() (R) gives quick stats.

• Utility: size/length (MATLAB); length, dim (R).

Scripts vs. Functions

• [!TIP] Use functions for reusable, testable code; use scripts for linear workflows or quick exploration.

2.3 Data Structures & Manipulation

MATLAB

R

• [!TIP] In R, [[ ]] extracts the element (drops wrapper), [ ] returns a sub-list/data frame. In MATLAB, () indexes cells, {} accesses contents.

2.4 Data Input/Output & Basic Preprocessing

Importing Data

• MATLAB: T = readtable('file.csv') (tabular); M = readmatrix('file.csv') (numeric matrix); importdata (flexible).

• R: df <- read.csv('file.csv'); df <- read.table('file.txt', sep='\t'); library(readr); df <- read_csv('file.csv') (faster, no factors by default).

Exporting Data

• MATLAB: writetable(T, 'out.csv'); writematrix(M, 'out.txt').

• R: write.csv(df, 'out.csv', row.names=FALSE).

Basic Inspection

• MATLAB: size(A), length(v), whos (list variables), class(x).

• R: dim(df), nrow(df), ncol(df), str(df) (structure), summary(df), head(df), tail(df).

Handling Missing Data

• Identify: isnan(A) (MATLAB); is.na(df) (R).

• Remove: rmmissing(A) (MATLAB); df[complete.cases(df), ] or na.omit(df) (R).

• Impute: Simple strategies: replace with mean/median (fillmissing in MATLAB; df$col[is.na(df$col)] <- mean(df$col, na.rm=TRUE) in R).

• [!CAUTION] na.omit() removes entire rows with any NA; may lose data. Consider imputation for critical features.

2.5 Applied Data Analysis & Visualization Basics

Descriptive Statistics

Basic Plotting

• MATLAB:

  • Line plot: plot(x,y).

  • Scatter: scatter(x,y).

  • Histogram: histogram(data).

  • Labels: xlabel('X'), ylabel('Y'), title('Title'), legend('label').

• R:

  • Base R: plot(x,y), hist(data), boxplot(data).

  • ggplot2: Grammar of graphics.

    
    library(ggplot2)
    
    ggplot(df, aes(x=var1, y=var2)) + geom_point() + labs(title="Title")
    
    

• [!TIP] In R, ggplot2 is layered and highly customizable; base R is quicker for simple plots.

Simple Data Transformations

• MATLAB: Arithmetic on arrays (A.*B element-wise); arrayfun(@func, A), cellfun(@func, C).

• R: Vectorized operations (x + 1); apply(df, 2, mean) (columns); lapply(lst, func) (list); sapply(lst, func) (simplified).

2.6 Debugging & Code Efficiency (Introductory)

Common Errors

• Syntax: Missing end, mismatched braces/parentheses.

• Dimension mismatch: A + B where sizes differ (MATLAB); cbind with unequal lengths (R).

• Type coercion: Implicit conversion (e.g., numeric to character in R vectors).

• Undefined variable: Typos or using variable before assignment.

• [!CAUTION] In R, $$\displaystyle ` indexing fails if column name has spaces/special chars; use backticks: `df $$column name`` or df[["column name"]].

Debugging Tools

• MATLAB:

  • Editor/Debugger: set breakpoints, step through.

  • dbstop if error: stops at error.

  • disp(), fprintf() for output.

• R:

  • traceback(): shows call stack after error.

  • debug(func): step into function.

  • browser(): pause execution.

  • print(), cat() for output.

Profiling & Efficiency

• MATLAB: profile on → run code → profile viewer; identifies slow functions.

• R: Rprof("prof.out") → code → Rprof(NULL); summaryRprof("prof.out").

• Pre-allocation (Critical for MATLAB):

  
  % Inefficient:
  
  A = [];
  
  for i=1:n
  
      A(i) = i^2;  % Grows array each iteration → slow
  
  end
  
  % Efficient:
  
  A = zeros(1,n);  % Pre-allocate
  
  for i=1:n
  
      A(i) = i^2;
  
  end
  
  

  • Time complexity remains $O(n)$ but constant factor is much lower.

  • In R, pre-allocate with vector() or numeric(n); lapply often avoids explicit loops.

• [!TIP] Exam Question: "Explain why pre-allocation improves performance in MATLAB." Answer: Avoids repeated memory allocation/copying during loop iterations.