Introduction To Statistical Decision Theory

Introduction to
statistical decision
theory
Group 7
Decision theory, in statistics, a set of quantitative methods for
reaching optimal decisions. A solvable decision problem must be
capable of being tightly formulated in terms of initial conditions
and choices or courses of action, with their consequences. In
general, such consequences are not known with
certainty but are expressed as a set
of probabilistic outcomes. Each
outcome is assigned a “utility” value
based on the preferences of the
decision maker. An
optimal decision, following the logic of the
theory, is one that maximizes the expected
utility.
Decision Making Under
Certainty
Making decisions under certainty is easy. The cause and
effect are known, and the risk involved is minimal. What’s
tough is making decisions under risk and uncertainty. The
outcome is unpredictable because you don’t have all the
information about the alternatives.
Before we learn deeper about decision-making under
risk and uncertainty, let’s look at each of these
situations.
Sometimes we have enough facts and evidence to know the

possible results of a decision. These are the most conducive
situations for decision-making because the outcomes are quite
obvious. For instance, if you drop a glass full of milk, the milk
Decision Making
under
Uncertainty
In case of an uncertain environment, you can’t predict the
outcomes as you have no information or data available. You have no control over what might
happen and don’t even know the options you have. It is like driving blindfolded where you know
you need to move but don’t know the type of vehicle or the road you will be taking. Such a
scenario will lead to decision-making under uncertainty. There are a variety of criteria that have
been proposed for the selection of an optimal course of action under the environment of
uncertainty. Each of these criteria make an assumption about the attitude of the decision-maker.
Maximin and maximax criteria, laplace, opportunity loss/ regret method, hurwicz method.
Decision Making under
Risk
Whenever the decision maker has some knowledge regarding the states of
nature, he/she may be able to assign subjective probability estimates for the
occurrence of each state. In such cases, the problem is classified as decision
making under risk. The decision-maker is able to assign probabilities based
on the occurrence of the states of nature. A problem of this kind arises
when the state of nature is unknown, but based on the objective or empirical
evidence, we can possibly assign probabilities to various states of nature. In
a number of problems on the basis of historical data and past experience,
we are able to assign probabilities to various states of nature. In such cases,
the pay-off matrix is of immense help for reaching an optimal decision by
assigning probabilities to various states of nature.
EMV Expected
Monetary
1) Acts: Value
There are many alternative courses of action in any
decision problem. But only some relevant alternatives
need be considered. For instance, the business firm may
decide to market its goods within the state or within the
country or beyond the boundaries of the country. Here,
there are three alternatives. There may be more such
alternatives. The final choice of any one will depend
upon the payoffs from each strategy.
2) States of Nature:
There are those possible events or the states of nature
which are uncertain but are vital for the choice of any
one of the alternative acts. For example, the radio dealer
does not
know how many radios he will be able to sell. There
is an element of uncertainty about it and for this
reason he cannot decide how many radios to buy.
This uncertainty is known as the state of nature or
the state of the world.
3) Outcomes:
There is an outcome of the combination of each of the
likely acts and possible states of nature. This is
otherwise known as conditional value. The outcome has
not much significant unless we calculate the pay-offs in
terms of monetary
gain or loss for each outcome. Thus outcome
refers to the result of the combination of an act
and each of the states of nature.
4) Pay-
off:
The pay-off deals with the monetary gain or loss from
each of the outcomes. It can be also in terms of cost-
saving or lime- saving but the expression of pay-off
should always be in quantitative terms to help precise
analysis. Therefore
where the value of output is expressed directly in
terms of gain expressed in money it is called pay-off.
The calculation of pay-off or utility of each outcome
has to be carefully done.
4) Expected Values of Each Act:
In practical business situation, there is risk and
uncertainty. In the case of risk, the probability of each
state of nature is known, and in uncertainty, it is
unknown. Therefore, each likely outcome of an act has to
be appraised with reference to the probability of
occurrence.
The expected value of a given act can be calculated by the

following formula:
Where P1 to Pn refers to event probabilities of events E1to En and

Oij, the pay-offs of the outcome with the combination of each
event and act. The expected value of each alternative is thus
calculated with reference to probability assigned to each state of

Introduction To Statistical Decision Theory

Uploaded by

Copyright:

Available Formats

Introduction To Statistical Decision Theory

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Introduction To Statistical Decision Theory

Uploaded by

Copyright:

Available Formats

Introduction to

Sometimes we have enough facts and evidence to know the

The expected value of a given act can be calculated by the

Where P1 to Pn refers to event probabilities of events E1to En and

You might also like