Results (
Thai) 1:
[Copy]Copied!
Decision trees can be used to predict the value of some response variable that has been identified as being important. Driven by the response variable, rules like “people who drink and smoke die before 70” can be found. Association rule learning aims at finding similar rules but now without focusing on a particular response variable. The goal is to find rules of the form IF X THEN Y where X is often called the antecedent and Y the consequent. Such rules are also denoted as X ⇒ Y . X and Y can be any conjunction of “variable = value” terms. The only requirement is that X and Y are nonempty and any variable appears at most once in X and Y . Examples are IF smoker = no AND age ≥ 70 THEN drinker = yes for Table 3.1 or IF logic ≤ 6 AND duration > 50 THEN result = failed for Table 3.2. Typically, only categorical variables are considered. However, there are various techniques to transform numerical variables in categorical ones.When learning association rules of the form X ⇒ Y , three metrics are frequentlyused: support, confidence, and lift. Let NX be the number of instances for whichX holds. NY is the number of instances for which Y holds. NX∧Y is the numberof instances for which both X and Y hold. N is the total number of instances. The support of a rule X ⇒ Y is defined as
Being translated, please wait..
