Unifying instance-based and rule-based induction

Domingos, Pedro

doi:10.1007/bf00058656

Cited by 119 publications

(131 citation statements)

References 40 publications

(67 reference statements)

Supporting

Mentioning

125

Contrasting

Unclassified

Order By: Relevance

“…The first 2/3 of the randomized data was reserved for training and/or cross-validation. For each algorithm, we report the average test set performance and sample standard deviation on 10 randomizations in each domain (Bay, 1999;De Groot, 1986;Domingos, 1996;Lim, Loh, & Shih, 2000). The best average test set performance was underlined and denoted in bold face for each domain.…”

Section: Description Of the Reference Algorithmsmentioning

confidence: 99%

See 1 more Smart Citation

Benchmarking Least Squares Support Vector Machine Classifiers

et al. 2004

View full text Add to dashboard Cite

Abstract. In Support Vector Machines (SVMs), the solution of the classification problem is characterized by a (convex) quadratic programming (QP) problem. In a modified version of SVMs, called Least Squares SVM classifiers (LS-SVMs), a least squares cost function is proposed so as to obtain a linear set of equations in the dual space. While the SVM classifier has a large margin interpretation, the LS-SVM formulation is related in this paper to a ridge regression approach for classification with binary targets and to Fisher's linear discriminant analysis in the feature space. Multiclass categorization problems are represented by a set of binary classifiers using different output coding schemes. While regularization is used to control the effective number of parameters of the LS-SVM classifier, the sparseness property of SVMs is lost due to the choice of the 2-norm. Sparseness can be imposed in a second stage by gradually pruning the support value spectrum and optimizing the hyperparameters during the sparse approximation procedure. In this paper, twenty public domain benchmark datasets are used to evaluate the test set performance of LS-SVM classifiers with linear, polynomial and radial basis function (RBF) kernels. Both the SVM and LS-SVM classifier with RBF kernel in combination with standard cross-validation procedures for hyperparameter selection achieve comparable test set performances. These SVM and LS-SVM performances are consistently very good when compared to a variety of methods described in the literature including decision tree based algorithms, statistical algorithms and instance based learning methods. We show on ten UCI datasets that the LS-SVM sparse approximation procedure can be successfully applied.

show abstract

Section: Description Of the Reference Algorithmsmentioning

confidence: 99%

“…Considering the Average Accuracy (AA) and Average Ranking (AR) over all domains (Bay, 1999;De Groot, 1986;Domingos, 1996), the RBF SVM gets the best average accuracy and the RBF LS-SVM yields the best average rank. There is no significant difference between the performance of both classifiers.…”

Section: Ls-svmmentioning

confidence: 99%

Benchmarking Least Squares Support Vector Machine Classifiers

et al. 2004

View full text Add to dashboard Cite

show abstract

“…However, this algorithm stores the entire data set in memory. Domingos also proposes 75 an integrated technique, the RISE algorithm, combining instance-based learning and rule induction [19]. Under 77 this algorithm, instances are treated as rules and data reduction is achieved using speciÿc rules formed by 79 generalization of instances.…”

Section: Motivation 21mentioning

confidence: 99%

“…conducts the ÿltering. 19 Unlike PGF1 which ÿlters on the original instances, PGF2 performs ÿltering on the prototype set. The pro-21 totype set usually contains intermediate prototypes and original instances.…”

Section: The Pgf Algorithmmentioning

confidence: 99%

Learning good prototypes for classification using filtering and abstraction of instances

Lam

Keung

Ling

2002

Pattern Recognition

View full text Add to dashboard Cite

“…Originally, the term was most often used to refer to the combination of analytical and inductive methods, but the combination of inductive methods having different biases is consistent with the term "multistrategy." This has been called "empirical multistrategy learning" (Domingos, 1996). We have also used the term "multistrategy learning" to refer to our framework in which individual learners are used as input to a combining method, which treats them as black boxes and attempts to model their behavior in pursuit of improved performance.…”

Section: Related Workmentioning

confidence: 99%

Untitled

Freitag¹

2000

Machine Learning

200

View full text Add to dashboard Cite

Abstract.We consider the problem of learning to perform information extraction in domains where linguistic processing is problematic, such as Usenet posts, email, and finger plan files. In place of syntactic and semantic information, other sources of information can be used, such as term frequency, typography, formatting, and mark-up. We describe four learning approaches to this problem, each drawn from a different paradigm: a rote learner, a term-space learner based on Naive Bayes, an approach using grammatical induction, and a relational rule learner. Experiments on 14 information extraction problems defined over four diverse document collections demonstrate the effectiveness of these approaches. Finally, we describe a multistrategy approach which combines these learners and yields performance competitive with or better than the best of them. This technique is modular and flexible, and could find application in other machine learning problems.

show abstract

Unifying instance-based and rule-based induction

Cited by 119 publications

References 40 publications

Benchmarking Least Squares Support Vector Machine Classifiers

Benchmarking Least Squares Support Vector Machine Classifiers

Learning good prototypes for classification using filtering and abstraction of instances

Untitled

Contact Info

Product

Resources

About