STAT2102 Basic Statistical Concepts and Methods II
2025 Spring Tutorial 2.
5th February, 2025
1 De Morgan’s Laws
Notation (basic set operations): Let A ∈ Ω and B ∈ Ω be two sets. The following are some common set
operations:
(i) Union: A ∪ B = {x : x ∈ A or x ∈ B}
(ii) Intersection: A ∩ B = {x : x ∈ A and x ∈ B}
(iii) Complement: Ac = {x : x ∈
/ A}
(iv) Difference: A \ B = {x : x ∈ A and x ∈
/ B}
(v) Symmetric difference: A△B = {x : x ∈ A or x ∈ B but not both}
The De Morgan’s Laws are two important rules useful for calculations. They are as follows:
(A ∪ B)c = Ac ∩ B c and (A ∩ B)c = Ac ∪ B c (1)
Example 1. Let Ω = [0, 10], A = [0, 5], B = [4, 6]. Then A ∪ B = [0, 6] and (A ∪ B)c = (6, 10]. Also,
Ac = (5, 10] and B c = [0, 4) ∪ (6, 10]. Therefore, Ac ∩ B c = (6, 10] ∩ ([0, 4) ∪ (6, 10]) = (6, 10]. Hence,
(A ∪ B)c = Ac ∩ B c .
Remark 1. Use Venn diagrams to help visualize the basic set operations and De Morgan’s Laws.
2 Taylor Expansion
Suppose f (x) is infinitely differentiable at x0 . Then the Taylor expansion of f (x) about x0 is given by:
f ′′ (x0 ) f (3) (x0 )
f (x) = f (x0 ) + f ′ (x0 )(x − x0 ) + (x − x0 )2 + (x − x0 )3 + · · · (2)
2! 3!
Example 2. Here are some Commonly used Taylor expansions at x0 = 0 (Maclaurin series):
x2 x3
(i) ex = 1 + x + 2! + 3! + ···
1
x3 x5
(ii) sin(x) = x − 3! + 5! − ···
x2 x4
(iii) cos(x) = 1 − 2! + 4! − ···
x2 x3
(iv) log(1 + x) = x − 2 + 3 − ···
Or alternatively, we can rewrite the Taylor expansion in a more compact form using the little o notation:
f ′′ (x0 ) f (n) (x0 )
f (x) = f (x0 ) + f ′ (x0 )(x − x0 ) + (x − x0 )2 + · · · + (x − x0 )n + o((x − x0 )n ) (3)
2! n!
The notation o((x − x0 )n ) denotes the error term in the Taylor expansion. It can be interpreted as a function
g(x)
g(x) such that limx→x0 (x−x 0)
n = 0 (But formally we should not use limit to define little o notation). Google
more about little o notation for more information, which is commonly used to describe the asymptotic
behavior in statistics and computer science.
√
Example 3. (i) an = n. limn→∞ ann = 0, so an = o(n).
(ii) an = nα , where 0 < α < 1. Then aα = o(n).
g(x)
(iii) g(x) = xn+1 . limx→0 xn = x, so g(x) = o(xn ).
3 Multiple Integrals
RR
Multiple integrals are useful for calculating probabilities. First consider a bivariate integrals D
1 dx dy,
where D = {(x, y) : 0 ≤ x ≤ y ≤ 1}. It can be computed as:
ZZ Z 1 Z 1 Z 1
1
1 dx dy = dx 1 dy = (1 − x) dx = (4)
D 0 x 0 2
A general example is computing the area of D = {(x1 , · · · , xn ) : 0 ≤ x1 ≤ · · · ≤ xn ≤ 1},
ZZ Z 1 Z xn Z x2
1
1 dx1 · · · dxn = dxn dxn−1 · · · dx1 = (5)
D 0 0 0 n!
Example 4 (Buffon’s Needle Problem). Suppose we have a floor with parallel lines drawn on it, separated
by a distance d. We drop a needle of length l(l ≤ d) on the floor. What is the probability that the needle
crosses a line?
Let x be the distance from the center of the needle to the nearest line, and θ be the angle between the needle
and the lines. Note that x and θ are uniformly distributed on [0, d2 ] and [0, π2 ] respectively. Moreover, they
4
are independent with joint density f (x, θ) = πd 1[0, d2 ] (x)1[0, π2 ] (θ). The needle crosses a line if x ≤ 2l sin(θ).
Therefore, the probability is:
Z π/2 Z l/2 sin(θ)
4 2l
P = dx dθ = (6)
θ=0 x=0 πd πd
