Minimax Error (decision theory problem)

In summary, the conversation discusses the difficulty of a homework problem in Decision Theory, specifically concerning the minimax criterion and finding the optimal decision point under a zero-one risk. The attempt at a solution involves maximizing the conditional Bayesian error and using derivatives to find the optimal decision point, but there are some concerns about the approach. The suggestion to consider expectimax is also mentioned.
  • #1
greatm31
1
0

Homework Statement



I've been trying to teach myself some Decision Theory using "Pattern Classification" by Duda, Heart, and Stork. It's a great book but the homework problems are giving me a lot of trouble. This one, concerning minimax, is particularly difficult:

"Assume we have one-dimensional Gaussian distributions [tex]p(x|\omega_{i})=N(\mu_{i},\sigma_{i}^{2})[/tex] for i=1,2, but completely unknown prior probabilities. Use the minimax criterion to find the optimal decision point x* in terms of [tex]\mu_{i}[/tex] and [tex]\sigma_{i}[/tex] under a zero-one risk."

Just to clarify, the zero-one risk means the loss function is 1 for an incorrect decision, and 0 for a correct decision.

2. The attempt at a solution

Now, my understanding of the minimax criterion is that we want to minimize the maximum Bayesian error for all possible priors. So I was thinking that the goal is to maximize the conditional Bayesian risk:

(key:
[tex]a_{i}[/tex] = action associated with selecting [tex]i[/tex]
[tex]\omega_{j}[/tex] = the so-called "state of nature" (tells whether [tex]j[/tex] is true)
[tex]x[/tex] = the data
[tex]c[/tex] = number of states of nature):

[tex]R(a_i|x) = \sum_{j}^c cost(a_{i}|\omega_{j}) \cdot P(\omega_{j}|x)[/tex]

Since the loss function is just 1 if [tex]i \neq j[/tex] and 0 if [tex]i=j[/tex],

[tex]R(a_{i}|x) = \sum_{j \neq i}^c P(\omega_{j}|x)[/tex]
[tex]= 1 - P(\omega_{i}|x)[/tex]

OK so this much is already developed in the book. I figure all I have to do is maximize the conditional error [tex]R(a_{i}|x)[/tex] and I'll have the "decision point" - the point at which you have the maximal Bayesian Risk, which should (I think) be independent of the prior probabilities [tex]P(\omega_{i})[/tex]. So that's what I did:

(here I'll use the abbreviation [tex]N_{i} = N(\mu_{i},\sigma_{i}^2)[/tex] = normal distribution with mean [tex]\mu_{i}[/tex] and variance [tex]\sigma_{i}^{2}[/tex]):

[tex]R(a_{i}|x) = 1 - P(\omega_{i}|x)[/tex]
[tex] = 1 - \frac{p(x|\omega_{i}) \cdot P(\omega_{i})}{p(x)}[/tex] (Bayes' Rule)
[tex] = 1 - \frac{N_{i} \cdot P(\omega_{i})}{N_{i} \cdot P(\omega_{i})+N_{j} \cdot P(\omega_{j})} [/tex]
[tex] = 1 - \frac{1}{1+\frac{P(\omega_{j})}{P(\omega_{i})} \cdot \frac{N_{j}}{N_{i}}} [/tex]

To minimize this, we maximize [tex]N_{j}/N_{i}[/tex] (thus independent of priors). I maximize this just by taking the first derivative and setting = 0:

[tex]0 = \frac{N_{i} \cdot N_{j}' - N_{j} \cdot N_{i}'}{N_{i}^2}[/tex]
[tex]\Rightarrow N_{i} \cdot N_{j}' = N_{j} \cdot N_{i}'[/tex]
[tex]\Rightarrow N_{i} \cdot \frac{-2 \cdot (x-\mu_{j})}{(2\sigma_{j}^{2})^{2}} \cdot N_{j} = N_{j} \cdot \frac{-2 \cdot (x-\mu_{i})}{(2\sigma_{i}^{2})^{2}} \cdot N_{i}[/tex] (deriv. of normal)
[tex]\Rightarrow \frac{x-\mu_{j}}{\sigma_{j}^{4}} = \frac{x-\mu_{i}}{\sigma_{i}^{4}}[/tex]
[tex]\Rightarrow x = \frac{\sigma_{i}^{4} \cdot \mu_{j} - \sigma_{j}^{4} \cdot \mu_{i}}{\sigma_{i}^4 - \sigma_{j}^{4}}[/tex]

So that's what I think x* is. The trouble is it doesn't make too much sense (what happens when [tex]\sigma_{i} = \sigma_{j}[/tex] ?). And also I'm not totally sure this is the right way of thinking about things.

Thanks for any help you can provide, even if it's just to tell me I'm completely wrong about my approach.
 
Physics news on Phys.org
  • #2
you might want to look at expectimax since you are dealing with probability
 
  • #3
I don't think that you use "the the minimax criterion" when solving this problem, so the answer cannot make sure that it is indepent of the prior probabilities...
 

What is Minimax Error and how is it related to decision theory?

Minimax Error is a concept in decision theory that refers to the minimum amount of error that can occur in a decision-making process. It is used to evaluate the performance of decision-making algorithms and models.

How is Minimax Error calculated?

Minimax Error is calculated by taking the maximum of the minimum possible error for each possible decision. This means that the algorithm or model is trying to minimize the maximum possible error that could occur in a decision-making scenario.

What is the difference between Minimax Error and other error measures?

Minimax Error differs from other error measures, such as mean squared error or mean absolute error, in that it focuses on the worst-case scenario. It is concerned with the maximum error that could occur, rather than the average error.

What are some real-world applications of Minimax Error?

Minimax Error is commonly used in decision-making processes in fields such as finance, economics, and game theory. It can also be applied in machine learning and artificial intelligence, where it is used to evaluate the performance of decision-making algorithms and models.

How can Minimax Error be minimized in decision-making processes?

Minimax Error can be minimized by using more accurate data, improving the decision-making algorithm or model, and reducing the number of possible decisions. It is also important to consider the potential consequences of each decision and make informed choices to minimize the maximum possible error.

Similar threads

  • Set Theory, Logic, Probability, Statistics
Replies
9
Views
1K
  • Engineering and Comp Sci Homework Help
Replies
1
Views
953
  • Thermodynamics
Replies
7
Views
1K
Replies
1
Views
791
Replies
2
Views
135
  • Quantum Physics
Replies
1
Views
233
  • Engineering and Comp Sci Homework Help
Replies
1
Views
1K
  • Precalculus Mathematics Homework Help
Replies
3
Views
2K
Replies
3
Views
1K
  • Set Theory, Logic, Probability, Statistics
Replies
1
Views
798
Back
Top