28.2. Examples of Linear Transformations

1. Identity Transformation: T(x) = x, ∀x ∈ R^n

The identity transformation is a function that returns the same input it receives. For any vector x in R^n, applying the identity transformation T just gives x back unchanged. This means if you take a vector and perform the identity transformation, there is no alteration to that vector.

Think of the identity transformation like a mirror: when you look into a mirror, you see a reflection of yourself exactly as you are—no changes. It's a straightforward scenario where the input equals the output.

2. Zero Transformation: T(x) = 0, ∀x ∈ R^n

The zero transformation is a function that maps every vector x in R^n to the zero vector, regardless of the input value. No matter what vector you provide to the transformation, the result is always a vector of all zeros. This illustrates a complete loss of information as everything maps to a single point (the zero vector).

Imagine pushing a ball down a hill. No matter how much push you consider for the ball's motion, it will always end up at the bottom (the last point) where it stops—equivalent to zero movement. Similarly, the zero transformation squashes every input down to the zero vector.

3. Scaling Transformation: T(x) = λx, λ ∈ R

A scaling transformation stretches or shrinks a vector by a scalar value λ. If λ is greater than 1, the vector's length increases (it gets longer), while if λ is between 0 and 1, the length reduces. If λ is negative, the direction of the vector also reverses. Essentially, it modifies the vector's magnitude and can change its direction based on the sign of λ.

Consider a zoom function on a camera: when you zoom in, objects appear enlarged (scaling up), and when you zoom out, they appear smaller (scaling down). Just like in the zoom function, the scaling transformation can change the size of vectors.

4. Rotation in R²: T = [x] [y] = [cosθ -sinθ] [x] [sinθ cosθ] [y]

A rotation transformation rotates points in the two-dimensional plane around the origin by a specified angle θ. The matrix defined here allows any vector (x, y) to be rotated to a new position (x', y') based on how much you rotate it around the origin. This transformation preserves distances and angles, meaning the vector stays the same length but shifts position.

Think about how a compass works: when you change the direction of the compass needle, it points in a different direction while remaining fixed at the same height above the ground. Similarly, rotation changes the direction of the vector (like the needle) without altering its length.

5. Projection onto a Line or Plane

Projection transformations involve mapping a vector onto a line or a plane. If you project vector x onto a line, you are finding the point on that line that is closest to x. This transformation effectively reduces the dimensionality of the vector (e.g., from 3D to 2D if projecting onto a plane), simplifying analyses and calculations in various fields.

If you shine a flashlight directly above a basketball, the light creates a circle on the ground beneath it—a projection of the ball's shadow. This shadow illustrates how the original object (the basketball) has been translated onto a different surface (the ground) without altering the nature of its outline.

--