DECISION RULES |
 |
What are THE DECISION (IF-THEN) RULES?
Decision Rules (If-then rules) are a method of data
formalization, used to
improve decision making. Decision Rules are universal, and can be used for
optimization of business processes as well as in scientific researches. The
representation of rules is very simple:
Decision rules are often providing the basis for so-called “expert systems”
which are widely used in many fields. The knowledge of experts, together with
newly discovered knowledge is encoded into the form of If-then rule set. When
exposed to the same or similar data, the expert system artificial intelligence
will perform in a similar manner to the expert.
Decision Rules Induction
Decision Rules may be induced from training data in two contrary methods:
- a bottom-up specific-to-general method
- a top-down general-to-specific method (like the decision trees method)
The initial state of decision rules induction is the collection of all training
examples. Each example can be a unique decision rule.
The aim of rules induction is to evolve the individual rules into general rules,
that will describe concerned classes, not only a single example.
The process of generalisation is iterative and is stopped when rules can not be
evolved any more, or because of reaching some stopping criteria.
How decision rules can be created with the help of
Estard Data Miner?
The
program can automatically induct decision (if-then) rules from the data
contained in data bases. The query settings for rules creation can be easily
formed with the help of a special query wizard. You can choose to create only
rules of a certain length or probability.
For each rule Estard Data Miner creates a set of features like probability,
cases, etc.
After creation, the rules can be viewed as a report or used for
various data analyses.
|
|
|
11.06.07
