<< Chapter < Page Chapter >> Page >

Review: pac bounds

Consider a finite collection of models F , and recall the basic PAC bound: for any δ > 0 , with probability at least 1 - δ

R ( f ) R ^ n ( f ) + log | F | + log ( 1 / δ ) 2 n , f F

where

R ^ n ( f ) = 1 n i = 1 n ( f ( X i ) , Y i ) R ( f ) = E ( f ( X ) , Y )

and the loss is assumed to be bounded between 0 and 1. Note that we can write the inequality above as:

R ( f ) R ^ n ( f ) + log | F | δ 2 n

Letting δ f = δ | F | , we have:

R ( f ) R ^ n ( f ) + log ( 1 / δ f ) 2 n

This is precisely the form of Hoeffding's inequality, with δ f in place of the usual δ . In effect, in order to have Hoeffding's inequality hold with probability 1 - δ for all f F , we must distribute the “ δ -budget” or “confidence-budget” over all f F (in this case, evenly distributed):

f F δ f = f F δ | F | = δ

However, to apply the union bound, we do not need to distribute δ evenly among the candidate models. We only require:

f F δ f = δ

So, if p ( f ) are positive numbers satisfying f F p ( f ) = 1 , then we can take δ f = p ( f ) δ . This provides two advantages:

  1. By choosing p ( f ) larger for certain f , we can preferentially treat those candidates
  2. We do not need F to be finite and we only require f F p ( f ) = 1

Prefix codes are one way to achieve this. If we assign a binary prefix code of length c ( f ) to each f F , then the values p ( f ) = 2 - c ( f ) satisfy f F p ( f ) 1 according to the Kraft inequality.

The main point of this lecture is to examine how PAC bounds of the form w.p. 1 - δ

R ( f ) R ^ n ( f ) + c ( f ) log 2 + log ( 1 / δ ) 2 n , f F

can be used to select a model that comes close to achieving the best possible performace

inf f F R ( f )

Let f ^ n be the model selected from F using the training data { X i , Y i } i = 1 n . We will specify this model in a moment, but keep in mind that it is notnecessarily the model with minimum empirical risk as before. We would like to have

E [ R ( f ^ n ) ] - inf f F R ( f )

as small as possible. First, for any δ > 0 , define

f ^ n δ = arg min f F R ^ n ( f ) + C ( f , n , δ )

where

C ( f , n , δ ) c ( f ) log 2 + log ( 1 / δ ) 2 n

Then w.p. 1 - δ

R ( f ) R ^ n ( f ) + C ( f , n , δ ) , f F

and in particular,

R ( f ^ n δ ) R ^ n ( f ^ n δ ) + C ( f ^ n δ , n , δ ) ,

so, by the definition of f ^ n δ , f F

R ( f ^ n δ ) R ^ n ( f ) + C ( f , n , δ ) .

We will make use of the inequality above in a moment. First note that f F

E [ R ( f ^ n δ ) ] - R ( f ) = E [ R ( f ^ n δ ) - R ^ n ( f ) ] + E [ R ^ n ( f ) - R ( f ) ]

The second term is exactly 0, since E [ R ^ n ( f ) ] = R ( f ) .

Now consider the first term E [ R ( f ^ n δ ) - R ^ n ( f ) ] . Let Ω be the set of events on which

R ( f ^ n δ ) R ^ n ( f ) - C ( f , n , δ ) , f F

From the bound above, we know that P ( Ω ) 1 - δ . Thus,

E [ R ( f ^ n δ ) - R ^ n ( f ) ] = E [ R ( f ^ n δ ) - R ^ n ( f ) | Ω ] P ( Ω ) + E [ R ( f ^ n δ ) - R ^ n ( f ) | Ω c ] ( 1 - P ( Ω ) ) C ( f , n , δ ) + δ ( since 0 R , R ^ 1 , P ( Ω ) 1 and 1 - P ( Ω ) δ ) = c ( f ) log 2 + log ( 1 / δ ) 2 n + δ = c ( f ) log 2 + 1 2 log n 2 n + 1 n ( by setting δ = 1 n )

We can summarize our analysis with the following theorem.

Theorem

Complexity regularized model selection

Let F be a countable collection of models, and assign a positive number c ( f ) to each f F such that f F 2 - c ( f ) 1 . Define the minimum complexity regularized risk model

f ^ n = arg min f F R ^ n ( f ) + c ( f ) log 2 + 1 2 log n 2 n

Then,

E [ R ( f ^ n ) ] inf f F R ( f ) + c ( f ) log 2 + 1 2 log n 2 n + 1 n

This shows that

R ^ n ( f ) + c ( f ) log 2 + 1 2 log n 2 n

is a reasonable surrogate for

R ( f ) + c ( f ) log 2 + 1 2 log n 2 n

Histogram classifiers

Let X = [ 0 , 1 ] d be the input space and Y = { 0 , 1 } be the output space. Let F k , k = 1, 2, ...  denotes the collection of histogram classification rules withk equal volume bins. One choice of prefix code for this example is: k = 1 code = 0 , k = 3 code = 10 , k = 3 code = 110 and so on .... Then, if first code is corresponding to k f F k , followed by k = log 2 | F k | bits to indicate which of the 2 k histogram rules in F k is under consideration, we have

f F k c ( f ) = 2 k b i t s

Let f ^ n be the model that solves the minimization i.e.,

min k 1 min f F k R ^ n ( f ) + 2 k log 2 + 1 2 log n 2 n

That is, for each k, let

f ^ n ( k ) = arg min f F k R ^ n ( f )

Then select the best k according to

k ^ = arg min k 1 R ^ n ( f ^ n ( k ) ) + 2 k log 2 + 1 2 log n 2 n

and set

f ^ n = f ^ n ( k ^ )

Then,

E [ R ( f ^ n ) ] inf k 1 min f F k R ( f ) + 2 k log 2 + 1 2 log n 2 n + 1 n

It is a simple exercise to show that if d = 2 and the Bayes decision boundary is a 1-d curve, then by setting k = n and selecting the best f from F n we have

E [ R ( f ^ n ) ] = O ( n - 1 / 4 )
The complexity regularized classifier f ^ n adaptively achieves this rate, without user intervention.

Questions & Answers

differentiate between demand and supply giving examples
Lambiv Reply
differentiated between demand and supply using examples
Lambiv
what is labour ?
Lambiv
how will I do?
Venny Reply
how is the graph works?I don't fully understand
Rezat Reply
information
Eliyee
devaluation
Eliyee
t
WARKISA
hi guys good evening to all
Lambiv
multiple choice question
Aster Reply
appreciation
Eliyee
explain perfect market
Lindiwe Reply
In economics, a perfect market refers to a theoretical construct where all participants have perfect information, goods are homogenous, there are no barriers to entry or exit, and prices are determined solely by supply and demand. It's an idealized model used for analysis,
Ezea
What is ceteris paribus?
Shukri Reply
other things being equal
AI-Robot
When MP₁ becomes negative, TP start to decline. Extuples Suppose that the short-run production function of certain cut-flower firm is given by: Q=4KL-0.6K2 - 0.112 • Where is quantity of cut flower produced, I is labour input and K is fixed capital input (K-5). Determine the average product of lab
Kelo
Extuples Suppose that the short-run production function of certain cut-flower firm is given by: Q=4KL-0.6K2 - 0.112 • Where is quantity of cut flower produced, I is labour input and K is fixed capital input (K-5). Determine the average product of labour (APL) and marginal product of labour (MPL)
Kelo
yes,thank you
Shukri
Can I ask you other question?
Shukri
what is monopoly mean?
Habtamu Reply
What is different between quantity demand and demand?
Shukri Reply
Quantity demanded refers to the specific amount of a good or service that consumers are willing and able to purchase at a give price and within a specific time period. Demand, on the other hand, is a broader concept that encompasses the entire relationship between price and quantity demanded
Ezea
ok
Shukri
how do you save a country economic situation when it's falling apart
Lilia Reply
what is the difference between economic growth and development
Fiker Reply
Economic growth as an increase in the production and consumption of goods and services within an economy.but Economic development as a broader concept that encompasses not only economic growth but also social & human well being.
Shukri
production function means
Jabir
What do you think is more important to focus on when considering inequality ?
Abdisa Reply
any question about economics?
Awais Reply
sir...I just want to ask one question... Define the term contract curve? if you are free please help me to find this answer 🙏
Asui
it is a curve that we get after connecting the pareto optimal combinations of two consumers after their mutually beneficial trade offs
Awais
thank you so much 👍 sir
Asui
In economics, the contract curve refers to the set of points in an Edgeworth box diagram where both parties involved in a trade cannot be made better off without making one of them worse off. It represents the Pareto efficient allocations of goods between two individuals or entities, where neither p
Cornelius
In economics, the contract curve refers to the set of points in an Edgeworth box diagram where both parties involved in a trade cannot be made better off without making one of them worse off. It represents the Pareto efficient allocations of goods between two individuals or entities,
Cornelius
Suppose a consumer consuming two commodities X and Y has The following utility function u=X0.4 Y0.6. If the price of the X and Y are 2 and 3 respectively and income Constraint is birr 50. A,Calculate quantities of x and y which maximize utility. B,Calculate value of Lagrange multiplier. C,Calculate quantities of X and Y consumed with a given price. D,alculate optimum level of output .
Feyisa Reply
Answer
Feyisa
c
Jabir
the market for lemon has 10 potential consumers, each having an individual demand curve p=101-10Qi, where p is price in dollar's per cup and Qi is the number of cups demanded per week by the i th consumer.Find the market demand curve using algebra. Draw an individual demand curve and the market dema
Gsbwnw Reply
suppose the production function is given by ( L, K)=L¼K¾.assuming capital is fixed find APL and MPL. consider the following short run production function:Q=6L²-0.4L³ a) find the value of L that maximizes output b)find the value of L that maximizes marginal product
Abdureman
types of unemployment
Yomi Reply
What is the difference between perfect competition and monopolistic competition?
Mohammed
Got questions? Join the online conversation and get instant answers!
Jobilize.com Reply

Get Jobilize Job Search Mobile App in your pocket Now!

Get it on Google Play Download on the App Store Now




Source:  OpenStax, Statistical learning theory. OpenStax CNX. Apr 10, 2009 Download for free at http://cnx.org/content/col10532/1.3
Google Play and the Google Play logo are trademarks of Google Inc.

Notification Switch

Would you like to follow the 'Statistical learning theory' conversation and receive update notifications?

Ask