next up previous
Link to Statistics and Actuarial Science main page
Next: Assignment 7 Up: 22S:193 Statistical Inference I Previous: Assignment 6

Solutions

2.27
a.
$ f(x)=\frac{1}{\sqrt{2\pi}}e^{-x^{2}/2}$ has a unique mode at $ x=0$.
b.
For

$\displaystyle f(x) = \begin{cases}1 & 0 \le x \le 1 0 & \text{otherwise} \end{cases}$    

any $ x \in [0,1]$ is a mode.
c.
Suppose, without loss of generality, that the point of symmetry is the origin. Suppose, without loss of generality, that the density is unimodal with mode $ a \ge 0$. For $ x \le y \le 0$ we have $ x \le y \le a$ and therefore by unimodality $ f(x) \le
f(y)$ and $ f(y) \le f(0)$. For $ 0 \le x \le y$ we have $ f(x)=f(-x)$ and $ f(y)=f(-y)$ by symmetry, and $ f(-y) \le f(-x)
\le f(0)$ by unimodality as before. So the point of symmetry 0 is a mode.
d.
The exponential distribution

$\displaystyle f(x) = \begin{cases}e^{-x} & x \ge 0 0 & x < 0 \end{cases}$    

is unimodal with mode $ x=0$ since $ f(x)$ is non-decreasing on $ (-\infty,0]$ and non-increasing (in fact strictly decreasing) on $ [0,\infty)$. If you switch the equality sign to the other side, it is not unimodal.

2.30
LATEX2Html is giving me grief about this one-it is available in the pdf version.
2.33
LATEX2Html is giving me grief about this one-it is available in the pdf version.
2.38
LATEX2Html is giving me grief about this one-it is available in the pdf version.


next up previous
Link to Statistics and Actuarial Science main page
Next: Assignment 7 Up: 22S:193 Statistical Inference I Previous: Assignment 6
Luke Tierney 2004-12-03