Representing classification problems in genetic programming

Loveard, Thomas; Ciesielski, Vic

doi:10.1109/cec.2001.934310

Cited by 135 publications

(128 citation statements)

References 3 publications

Supporting

Mentioning

125

Contrasting

Unclassified

Order By: Relevance

“…GP research has considered a variety of kinds of classifier programs, using different program representations, including decision tree classifiers, classification rule sets [6], and linear and graph classifiers [4]. Recently, a new form of classifier representation -numeric expression (tree-like) classifiers -has been developed using GP [7,8,9,10]. This form has been successfully applied to real world classification problems such as detecting and recognising particular classes of objects in images [8,9,11,12], demonstrating the potential of GP as a general method to solve classification problems.…”

Section: Introductionmentioning

confidence: 99%

“…For the simple binary classification case, this translation can be based on the sign of the numeric value [8,7,9,13,14,15,16]; for multiclass problems, finding the appropriate boundary values to separate the different classes is more difficult.…”

Section: Introductionmentioning

confidence: 99%

“…This includes a primary static method such as object classification map or static range selection [10,7,12], dynamic range selection [7], centred and slotted dynamic class boundary determination methods [15,16]. Past work has demonstrated the effectiveness of these approaches, particularly the dynamic methods, on a number of object classification problems.…”

mentioning

confidence: 99%

See 2 more Smart Citations

Using Gaussian distribution to construct fitness functions in genetic programming for multiclass object classification

Zhang

Smart

2006

Pattern Recognition Letters

View full text Add to dashboard Cite

This paper describes a new approach to the use of Gaussian distribution in genetic programming (GP) for multiclass object classification problems. Instead of using predefined multiple thresholds to form different regions in the program output space for different classes, this approach uses probabilities of different classes, derived from Gaussian distributions, to construct the fitness function for classification. Two fitness measures, overlap area and weighted distribution distance, have been developed. Rather than using the best evolved program in a population, this approach uses multiple programs and a voting strategy to perform classification. The approach is examined on three multiclass object classification problems of increasing difficulty and compared with a basic GP approach. The results suggest that the new approach is more effective and more efficient than the basic GP approach. Although developed for object classification, this approach is expected to be able to be applied to other classification problems.Keywords Probability based genetic programming, object recognition, object detection, fitness function, multiclass classification Author InformationThe first author is a staff member in computer science, Victoria University of Wellington, New Zealand. The second author is postgraduate student in computer science, Victoria University of Wellington, New Zealand. Using Gaussian Distribution to Construct Fitness Functions in Genetic Programming AbstractThis paper describes a new approach to the use of Gaussian distribution in genetic programming (GP) for multiclass object classification problems. Instead of using predefined multiple thresholds to form different regions in the program output space for different classes, this approach uses probabilities of different classes, derived from Gaussian distributions, to construct the fitness function for classification. Two fitness measures, overlap area and weighted distribution distance, have been developed. Rather than using the best evolved program in a population, this approach uses multiple programs and a voting strategy to perform classification. The approach is examined on three multiclass object classification problems of increasing difficulty and compared with a basic GP approach. The results suggest that the new approach is more effective and more efficient than the basic GP approach. Although developed for object classification, this approach is expected to be able to be applied to other classification problems.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Using Gaussian distribution to construct fitness functions in genetic programming for multiclass object classification

Zhang

Smart

2006

Pattern Recognition Letters

View full text Add to dashboard Cite

show abstract

“…Recently, a new form of classifier representation -numeric expression classifiers -has been developed using GP [4][5][6][7]. In these years, this form has become the "standard form" of GP and has been successfully applied to some real world classification problems such as detecting and recognising particular classes of objects in images [5,6,8,9], demonstrating the potential of GP as a general method for classification problems.…”

Section: Introductionmentioning

confidence: 99%

Multiclass Object Classification Using Genetic Programming

Zhang

Smart

2004

Lecture Notes in Computer Science

View full text Add to dashboard Cite

We describe an approach to the use of genetic programming for multi-class object classification problems. Rather than using fixed static thresholds as boundaries to distinguish between different classes, this approach introduces two methods of classification where the boundaries between different classes can be dynamically determined during the evolutionary process. The two methods are centred dynamic class boundary determination and slotted dynamic class boundary determination. The two methods are tested on four object classification problems of increasing difficulty and are compared with the commonly used static class boundary method. The results suggest that, while the static class boundary method works well on relatively easy object classification problems, the two dynamic class boundary determination methods outperform the static method for more difficult, multiple class object classification problems.Keywords Genetic programming, genetic algorithms, dynamic class boundary determination, object recognition. Author InformationMengjie Zhang is an academic staff member in computer science and Will Smart is a postgraduate student in computer science. Both authors are in the School of Mathematical and Computing Sciences, Victoria University of Wellington, New Zealand. Multiclass Object Classification Using Genetic Programming Mengjie Zhang and Will SmartSchool of Mathematical and Computing Sciences Victoria University of Wellington, P. O. Box 600, Wellington, New Zealand, Email: {mengjie,willsmart}@mcs.vuw.ac.nz Abstract. We describe an approach to the use of genetic programming for multi-class object classification problems. Rather than using fixed static thresholds as boundaries to distinguish between different classes, this approach introduces two methods of classification where the boundaries between different classes can be dynamically determined during the evolutionary process. The two methods are centred dynamic class boundary determination and slotted dynamic class boundary determination. The two methods are tested on four object classification problems of increasing difficulty and are compared with the commonly used static class boundary method. The results suggest that, while the static class boundary method works well on relatively easy, linearly separable object classification problems, the two dynamic class boundary determination methods outperform the static method for more difficult, multiple class object classification problems.

show abstract

“…However the small difference between the two GP approaches might also be due to using one tree instead of three. [Loveard and Ciesielski, 2001] investigates other ways to evolve multi-class classifiers. Figure 1 shows the exponential nature (common in many chemical systems) of the distribution of IC50.…”

Section: Discussionmentioning

confidence: 99%

Predicting biochemical interactions - human P450 2D6 enzyme inhibition

Langdon

Barrett²,

Buxton³

The 2003 Congress on Evolutionary Computation, 2003. CEC '03.

View full text Add to dashboard Cite

Abstract-In silico screening of chemical libraries or virtual chemicals may reduce drug discovery and medicine optimisation lead times and increase the probability of success by directing search through chemical space. About a dozen intelligent pharmaceutical QSAR modelling techniques were used to predict IC50 concentration (three classes) of drug interaction with a cell wall enzyme (P450 CYC2D6). Genetic programming gave comprehensible cheminformatics models which generalised best. This was shown by a blind test on GlaxoWelcome molecules of machine learning knowledge nuggets mined from SmithKline Beecham compounds. Performance on similar chemicals (interpolation) and diverse chemicals (extrapolation) suggest generalisation is more difficult than avoiding over fitting.Two GP approaches, classification via regression using a multi-objective fitness measure and a direct winner takes all (WTA) or one versus all (OVA) classification, are described. Predictive rules were compressed by separate follow up GP runs seeded with the best program.

show abstract

Representing classification problems in genetic programming

Cited by 135 publications

References 3 publications

Using Gaussian distribution to construct fitness functions in genetic programming for multiclass object classification

Using Gaussian distribution to construct fitness functions in genetic programming for multiclass object classification

Multiclass Object Classification Using Genetic Programming

Predicting biochemical interactions - human P450 2D6 enzyme inhibition

Contact Info

Product

Resources

About