Mining multiple comprehensible classification rules using genetic programming

Tan, Kay Chen; Tay, Arthur; Lee, T.H.; Heng, Chun-Huat

doi:10.1109/cec.2002.1004431

Cited by 42 publications

(28 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…A GP classifier has been proposed in [41], using an artificial immune system-like memory vector to generate multiple comprehensible classification rules. It introduces a concept mapping technique to evaluate the quality of the rules.…”

Section: Related Workmentioning

confidence: 99%

Opposition-Based Discrete PSO Using Natural Encoding for Classification Rule Discovery

Khan

Baig

Iqbal

2012

International Journal of Advanced Robotic Systems

View full text Add to dashboard Cite

In this paper we present a new Discrete Particle Swarm Optimization approach to induce rules from discrete data. The proposed algorithm, called Oppositionbased Natural Discrete PSO (ONDPSO), initializes its population by taking into account the discrete nature of the data. Particles are encoded using a Natural Encoding scheme. Each member of the population updates its position iteratively on the basis of a newly designed position update rule. Opposition-based learning is implemented in the optimization process. The encoding scheme and position update rule used by the algorithm allows individual terms corresponding to different attributes within the rule's antecedent to be a disjunction of the values of those attributes. The performance of the proposed algorithm is evaluated against seven different datasets using a tenfold testing scheme. The achieved median accuracy is compared against various evolutionary and non-evolutionary classification techniques. The algorithm produces promising results by creating highly accurate and precise rules for each dataset.

show abstract

Section: Related Workmentioning

confidence: 99%

Opposition-Based Discrete PSO Using Natural Encoding for Classification Rule Discovery

Khan

Baig

Iqbal

2012

International Journal of Advanced Robotic Systems

View full text Add to dashboard Cite

show abstract

“…A classifier system based on the Michigan representation is used in [30], where several classification rules are evolved. Each rule consists of enumerations of attribute value pairs, combined with only two Boolean functions, AND and NOT.…”

Section: Extensions Of Standard Gp-classificationmentioning

confidence: 99%

Improving GP classification performance by injection of decision trees

König

Johansson

Löfström

et al. 2010

IEEE Congress on Evolutionary Computation

View full text Add to dashboard Cite

Abstract-This paper presents a novel hybrid method combining genetic programming and decision tree learning. The method starts by estimating a benchmark level of reasonable accuracy, based on decision tree performance on bootstrap samples of the training set. Next, a normal GP evolution is started with the aim of producing an accurate GP. At even intervals, the best GP in the population is evaluated against the accuracy benchmark. If the GP has higher accuracy than the benchmark, the evolution continues normally until the maximum number of generations is reached. If the accuracy is lower than the benchmark, two things happen. First, the fitness function is modified to allow larger GPs, able to represent more complex models. Secondly, a decision tree with increased size and trained on a bootstrap of the training data is injected into the population. The experiments show that the hybrid solution of injecting decision trees into a GP population gives synergetic effects producing results that are better than using either technique separately. The results, from 18 UCI data sets, show that the proposed method clearly outperforms normal GP, and is significantly better than the standard decision tree algorithm.

show abstract

“…There is often no single rule that can cover all instances of a class: hence there is a need to discover more rules that can predict all instances of a class, but at the same time do not overlap with instances of another class. The CORE also applied the principle of controlling the fitness of rules via the concept of minimum support as proposed by Tan et al (2002a). Every instance in a dataset is called a token, for which all chromosomes in the population will compete to capture.…”

Section: Token Competitionmentioning

confidence: 99%

“…Data mining is an automated process of extracting structured knowledge from databases, which is often referred to as an essential step in the overall process of discovering useful knowledge from data, called knowledge discovery from database (KDD) (Fayyad 1997, Liu andMotoda 1998). In recent years, there have been numerous attempts to apply evolutionary computation techniques in data mining to tackle the problem of knowledge extraction and classification (Hruschka and Ebecken 2000, Wong and Leung 2000, Tan et al 2002a or to accomplish tasks in different domains (Banzhaf et al 1998, Cattral et al 1999, Pozo and Hasse 2000, Brameier and Banzhaf 2001.…”

Section: Introductionmentioning

confidence: 99%