User:IssaRice/Little o notation: Difference between revisions

From Machinelearning
(Created page with " Can we write just <math>f \in o(g)</math> or <math>f = o(g)</math> or <math>f(x) \in o(g(x))</math> or <math>f(x) = o(g(x))</math>? In general we can't because for this not...")
 
 
(26 intermediate revisions by the same user not shown)
Line 1: Line 1:
==Definition==


'''Definition''' (little o near a point). Let <math>f : \mathbf R \to \mathbf R</math> and <math>g : \mathbf R \to \mathbf R</math> be two functions, and let <math>a \in \mathbf R</math>. We say that <math>f</math> is little o of <math>g</math> near <math>a</math> iff for every <math>\epsilon > 0</math> there exists <math>\delta > 0</math> such that <math>|x - a| < \delta</math> implies <math>|f(x)| < \epsilon|g(x)|</math>. Some equivalent ways to say the same thing are:


{| class="wikitable"
|-
! Notation !! Comments
|-
| <math>f</math> is little o of <math>g</math> near <math>a</math> || This is a point-free notation.
|-
| <math>f(x) \in o(g(x))</math> as <math>x \to a</math> || This is in point notation, as the variable <math>x</math> appears in the notation. This allows us to define functions anonymously. For example, we can say <math>x^2 \in o(x)</math> as <math>x \to 0</math>; we didn't even name the functions. As the appearance of the symbol "<math>\in</math>" suggests, in this notation we think of <math>o(g(x))</math> as a set, namely the set of all functions that are <math>o(g(x))</math> as <math>x \to a</math>.
|-
| <math>f(x) = o(g(x))</math> as <math>x \to a</math> || This is in point notation. As the equality symbol suggests, in this notation we think of f as a concrete manifestation of a function that is <math>o(g(x))</math> near <math>a</math>. This allows us to algebraically manipulate the expression <math>o(g(x))</math> along with all our other expressions.
|-
| <math>f \in o(g)</math> near <math>a</math> || This is a point-free notation. As the appearance of the symbol "<math>\in</math>" suggests, in this notation we think of <math>o(g)</math> as a set, namely the set of all functions that are <math>o(g)</math> near <math>a</math>. In other words, <math>o(g) = \{f : \forall \epsilon > 0 \ \exists \delta > 0\ \forall x\ (|x-a| < \delta \implies |f(x)| < \epsilon|g(x)|)\}</math>
|-
| <math>f = o(g)</math> near <math>a</math> || This is a point-free notation.
|}


Can we write just <math>f \in o(g)</math> or <math>f = o(g)</math> or <math>f(x) \in o(g(x))</math> or <math>f(x) = o(g(x))</math>? In general we can't because for this notation to make sense, we also need to know where the argument <math>x</math> is going. In algorithms, we have <math>x \to \infty</math>, but in analysis (e.g. in some definitions of differentiability) we have <math>x \to 0</math>
'''Definition''' (little o at infinity). Let <math>f : \mathbf R \to \mathbf R</math> and <math>g : \mathbf R \to \mathbf R</math> be two functions. We say that <math>f</math> is little o of <math>g</math> at positive infinity (or equivalently <math>f(x)</math> is little of <math>g(x)</math> as <math>x \to \infty</math>) iff for every real <math>\epsilon > 0</math> there exists a real number <math>M</math> such that for all <math>x</math>, if <math>x > M</math> then <math>|f(x)| < \epsilon|g(x)|</math>. We say that <math>f</math> is little o of <math>g</math> at negative infinity (or equivalently <math>f(x)</math> is little of <math>g(x)</math> as <math>x \to -\infty</math>) iff for every real <math>\epsilon > 0</math> there exists a real number <math>M</math> such that for all <math>x</math>, if <math>x < M</math> then <math>|f(x)| < \epsilon|g(x)|</math>.
 
'''Exercise'''. Show that in the definition of little o at positive infinity, "there exists a real number <math>M</math>" can be replaced by "there exists a real number <math>M > 0</math>". Show that in the definition of little o at negative infinity, "there exists a real number <math>M</math>" can be replaced by "there exists a real number <math>M < 0</math>".
 
'''Exercise'''. Can we write just <math>f \in o(g)</math> or <math>f = o(g)</math> or <math>f(x) \in o(g(x))</math> or <math>f(x) = o(g(x))</math>?
 
{{collapsible solution|In general we can't because for this notation to make sense, we also need to know where the argument <math>x</math> is going. In algorithms, we have <math>x \to \infty</math>, but in analysis (e.g. in some definitions of differentiability) we have <math>x \to 0</math>.}}
 
'''Exercise'''. If we are being a little pedantic, what is wrong with saying "<math>f \in o(g)</math> as <math>x \to a</math>"?
 
{{collapsible solution|We are saying <math>x \to a</math>, but we haven't clarified what <math>x</math> is. Instead, we are relying on the reader to assume that <math>x</math> is an argument to <math>f</math> and <math>g</math>.}}
 
'''Exercise'''. Interpret the meaning of <math>x^2 \in o(x)</math>.
 
{{collapsible solution|It depends on where <math>x</math> is going. We want <math>|x^2| < \epsilon |x|</math> whenever <math>|x-a|<\delta</math>, so this is only true when <math>a = 0</math>.}}
 
==Properties==
 
'''Proposition'''. Let <math>f : \mathbf R \to \mathbf R</math> and <math>g : \mathbf R \to \mathbf R</math> be two functions, and suppose <math>g(x) \ne 0</math> for all <math>x \in \mathbf R</math>. Then f is little o of g near a if and only if <math>\lim_{x\to a} \frac{f(x)}{g(x)} = 0</math>.
 
'''Proposition'''. transitivity
 
'''Proposition'''. we can replace the <math><</math> in the definition with <math>\leq</math>, right?
 
'''Exercise'''. Let <math>c, x_0 \in \mathbf R</math> be constants. Interpret the statement "<math>o(c(x-x_0) + o(x-x_0)) \in o(x-x_0)</math> as <math>x \to x_0</math>".
 
{{collapsible solution|
TODO: be careful with universal vs existential quantifiers.
 
The statement is saying <math>f(x) \in o(x-x_0)</math> where <math>f</math> is some function such that <math>\lim_{x\to x_0} \frac{f(x)}{c(x-x_0) + o(x-x_0)} = 0</math>.
 
Because of the nested little o, we need to expand again and introduce <math>g</math>, where <math>g(x) \in o(x-x_0)</math> so <math>\lim_{x\to x_0} \frac{g(x)}{x-x_0} = 0</math>.
 
Now we verify:
 
<math>\lim_{x\to x_0} \frac{f(x)}{x-x_0} = \lim \frac{f(x)}{c(x-x_0) + g(x)} \lim \frac{c(x-x_0) + g(x)}{x-x_0} = 0 \cdot (c + 0) = 0</math>
 
Could <math>g</math> have been arbitrary? In other words, could we have said <math>o(c(x-x_0) + o(h(x))) \in o(x-x_0)</math> for arbitrary <math>h(x)</math>? To compute the limit <math>\lim_{x\to x_0} \frac{f(x)}{x-x_0}</math> we actually used the limit laws, which require that the right hand limit exist. This means that we needed <math>\lim_{x\to x_0} g(x)/h(x)</math> to exist.}}
 
==References==
 
<ref>https://sites.math.washington.edu/~folland/Math134/lin-approx.pdf</ref>
 
<ref>https://math.stackexchange.com/a/1784280/35525</ref>
 
<references/>

Latest revision as of 17:29, 29 November 2018

Definition

Definition (little o near a point). Let and be two functions, and let . We say that is little o of near iff for every there exists such that implies . Some equivalent ways to say the same thing are:

Notation Comments
is little o of near This is a point-free notation.
as This is in point notation, as the variable appears in the notation. This allows us to define functions anonymously. For example, we can say as ; we didn't even name the functions. As the appearance of the symbol "" suggests, in this notation we think of as a set, namely the set of all functions that are as .
as This is in point notation. As the equality symbol suggests, in this notation we think of f as a concrete manifestation of a function that is near . This allows us to algebraically manipulate the expression along with all our other expressions.
near This is a point-free notation. As the appearance of the symbol "" suggests, in this notation we think of as a set, namely the set of all functions that are near . In other words,
near This is a point-free notation.

Definition (little o at infinity). Let and be two functions. We say that is little o of at positive infinity (or equivalently is little of as ) iff for every real there exists a real number such that for all , if then . We say that is little o of at negative infinity (or equivalently is little of as ) iff for every real there exists a real number such that for all , if then .

Exercise. Show that in the definition of little o at positive infinity, "there exists a real number " can be replaced by "there exists a real number ". Show that in the definition of little o at negative infinity, "there exists a real number " can be replaced by "there exists a real number ".

Exercise. Can we write just or or or ?

Expand to see solution:

In general we can't because for this notation to make sense, we also need to know where the argument is going. In algorithms, we have , but in analysis (e.g. in some definitions of differentiability) we have .

Exercise. If we are being a little pedantic, what is wrong with saying " as "?

Expand to see solution:

We are saying , but we haven't clarified what is. Instead, we are relying on the reader to assume that is an argument to and .

Exercise. Interpret the meaning of .

Expand to see solution:

It depends on where is going. We want whenever , so this is only true when .

Properties

Proposition. Let and be two functions, and suppose for all . Then f is little o of g near a if and only if .

Proposition. transitivity

Proposition. we can replace the in the definition with , right?

Exercise. Let be constants. Interpret the statement " as ".

Expand to see solution:

TODO: be careful with universal vs existential quantifiers.

The statement is saying where is some function such that .

Because of the nested little o, we need to expand again and introduce , where so .

Now we verify:

Could have been arbitrary? In other words, could we have said for arbitrary ? To compute the limit we actually used the limit laws, which require that the right hand limit exist. This means that we needed to exist.

References

[1]

[2]