Cumulative distribution functions

Cumulative distribution functions

Spec SD6 · AQA A-Level Further Maths: Revision notes

Studying this course? Keep your progress in one place.

Key facts

  • The cdf is F(x)=P(X≤x)=∫−∞xf(t) dtF(x)=P(X\le x)=\int_{-\infty}^xf(t)\,dt.
  • FF is non-decreasing, runs from 00 to 11, and has no jumps for a continuous variable.
  • f(x)=F′(x)f(x)=F'(x) for each piece.
  • P(a<X<b)=F(b)−F(a)P(a<X<b)=F(b)-F(a) and the median solves F(m)=12F(m)=\frac12.
  • For a cdf in pieces, use FF at the join to choose the right piece.

What the cdf is

F(x)F(x) is the probability that XX is at most xx: the area under the pdf up to xx.

The cumulative distribution function (cdf) of a continuous random variable XX is

F(x)=P(X≤x)=∫−∞xf(t) dt.F(x)=P(X\le x)=\int_{-\infty}^{x}f(t)\,dt.

Properties: FF is non-decreasing; F(x)→0F(x)\to0 as x→−∞x\to-\infty and F(x)→1F(x)\to1 as x→∞x\to\infty, so 0≤F(x)≤10\le F(x)\le1; for a continuous variable FF has no jumps.

11.522.533.54−0.20.20.40.60.811.2xyy = 1(1, 0)(2, 0.875)F(x) = 1 − x⁻³
The cdf F(x) = 1 − x⁻³ for x ≥ 1 rises from 0 and levels off at 1.

Which cannot be a value of F(x)F(x)?

From the pdf to the cdf

Integrate ff from the bottom of the range up to xx, using a dummy variable tt.

Integrate ff from the bottom of its range up to xx, using a dummy variable tt:

F(x)=∫axf(t) dtfor a≤x≤b,F(x)=\int_{a}^{x}f(t)\,dt\quad\text{for }a\le x\le b,

with F(x)=0F(x)=0 below the range and F(x)=1F(x)=1 above it.

For a pdf in pieces, add the total area of the earlier pieces to the integral over the current piece, so FF is continuous at each join.

0.20.40.60.811.21.41.61.820.20.40.60.81xyF(1) = 0.5F(2) = 1f(x)F(x)
f(x) = (6x − 3x²)/4 and F(x) = (3x² − x³)/4 on 0 ≤ x ≤ 2: F(x) is the area under f up to x, so F(1) = 0.5
  • F(x)F(x) within the range∫axf(t) dt\int_a^xf(t)\,dt
  • Below / above the range00 / 11

Worked example

f(t)=3t−4f(t)=3t^{-4} for t≥1t\ge1. Find F(x)F(x).

f(t)=t2f(t)=\frac{t}{2} on [0,2][0,2]. What is F(x)F(x) on that range?

From the cdf to the pdf

Differentiate each piece: f(x)=dFdxf(x)=\frac{dF}{dx}.

Differentiate each piece:

f(x)=dFdx.f(x)=\frac{dF}{dx}.

Outside the range f(x)=0f(x)=0.

If F(x)=3x2−x34F(x)=\frac{3x^2-x^3}{4} on 0≤x≤20\le x\le2, then f(x)=6x−3x24f(x)=\frac{6x-3x^2}{4} on that range.

An unknown constant in FF is found from continuity or from F=1F=1 at the top of the range: if F(x)=k(x2+2x)F(x)=k(x^2+2x) on 0≤x≤20\le x\le2 then F(2)=8k=1F(2)=8k=1, so k=18k=\frac18.

0.20.40.60.811.21.41.61.822.22.40.20.40.60.811.2xyF(2) = 1F(x) = (x² + 2x)/8
F(x) = k(x² + 2x) on 0 ≤ x ≤ 2: F must reach 1 at the top of the range, so 8k = 1 and k = 1/8
  • pdf from cdff(x)=F′(x)f(x)=F'(x)
  • Unknown constantuse F=1F=1 at the top of the range

F(x)=x29F(x)=\frac{x^2}{9} on 0≤x≤30\le x\le3. What is f(x)f(x) on that range?

Using the cdf

Probabilities are differences of FF; the median solves F(m)=12F(m)=\frac12.

Probabilities come straight from FF, with no integration:

P(a<X<b)=F(b)−F(a),P(X>a)=1−F(a).P(a<X<b)=F(b)-F(a),\qquad P(X>a)=1-F(a).

The median mm satisfies F(m)=12F(m)=\frac12; the quartiles satisfy F(Q1)=14F(Q_1)=\frac14 and F(Q3)=34F(Q_3)=\frac34.

When FF is in pieces, compare the target with FF at the join to choose which piece to solve in, then check the root lies in that piece.

0.20.40.60.811.21.41.61.820.20.40.60.81xyF = ½median m = 1F(0.5)F(1.5)F(x)
F(x) = (3x² − x³)/4: the median is where F = ½, and P(a < X < b) = F(b) − F(a)
  • P(a<X<b)P(a<X<b)F(b)−F(a)F(b)-F(a)
  • MedianF(m)=12F(m)=\frac12

Worked example

F(x)=x26F(x)=\frac{x^2}{6} for 0≤x≤20\le x\le2 and x3\frac x3 for 2<x≤32<x\le3. Find P(1<X<2.5)P(1<X<2.5), the median and the upper quartile.

F(x)=x4F(x)=\frac{x}{4} on 0≤x≤40\le x\le4. What is P(1<X<3)P(1<X<3)?

Try an exam question

A continuous random variable XX has cumulative distribution function F(x)=x26F(x)=\dfrac{x^2}{6} for 0≤x≤20\le x\le2, F(x)=x3F(x)=\dfrac x3 for 2<x≤32<x\le3, F(x)=0F(x)=0 for x<0x<0 and F(x)=1F(x)=1 for x>3x>3. Find P(1<X<2.5)P(1<X<2.5) and the median of XX.

[5 marks]

That's the notes covered.

Carry on to the next subtopic.

Try a question

Find P(X≤1)P(X\le1).

1 mark, marked against the mark scheme

Revision notes