Associative Classification

Associative Classification

Associative classification

o   Association rules are generated and analyzed for use in classification

o   Search for strong associations between frequent patterns (conjunctions of attribute-value pairs) and class labels

o   Classification: Based on evaluating a set of rules in the form of

P₁ ^ p₂ … ^ p_l ^à ―A_class = C‖ (conf, sup)

Why effective?

o   It explores highly confident associations among multiple attributes and may overcome some constraints introduced by decision-tree induction, which considers only one attribute at a time

In many studies, associative classification has been found to be more accurate than some traditional classification methods, such as C4.

Typical Associative Classification Methods

CBA (Classification By Association: Liu, Hsu & Ma, KDD’98)

o   Mine association possible rules in the form of

^o     Cond-set (a set of attribute-value pairs) ^à class label

o   Build classifier: Organize rules according to decreasing precedence based on confidence and then support

CMAR (Classification based on Multiple Association Rules: Li, Han, Pei, ICDM’01)

o   Classification: Statistical analysis on multiple rules

CPAR (Classification based on Predictive Association Rules: Yin & Han, SDM’03)

o   Generation of predictive rules (FOIL-like analysis)

o   High efficiency, accuracy similar to CMAR

RCBT  (Mining  top-k  covering  rule  groups  for  gene  expression  data,  Cong  et  al.

SIGMOD’05)

o   Explore high-dimensional classification, using top-k rule groups

o   Achieve high classification accuracy and high run-time efficiency