Combining Classifiers: Soft Computing Solutions

Kuncheva, Ludmila I.; Pal, Souvik; Pal, Abhijit

doi:10.1142/9789812386533_0015

Cited by 94 publications

(58 citation statements)

References 58 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…They propose to change (i) starting point in the hypothesis space, (ii) training data, (iii) classifier architectures, and (iv) traversal of the hypothesis space. In an early work, Kuncheva [20] proposes four different methods to build a multiclassifier system: (i) using different combiners when already trained base classifiers are given, (ii) using different algorithms and starting parameters, (iii) using different feature subsets, and (iv) using different training sets.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Eigenclassifiers for combining correlated classifiers

Ulas

Yıldız

Alpaydın

2012

Information Sciences

View full text Add to dashboard Cite

Section: Related Workmentioning

confidence: 99%

“…BAG: The original Bagging algorithm proposed in [7]. We train decision tree ensembles of size 5,10,15,20,25,30 and choose the one with the best val-B accuracy.…”

Section: Comparison With Adaboost and Baggingmentioning

confidence: 99%

Eigenclassifiers for combining correlated classifiers

Ulas

Yıldız

Alpaydın

2012

Information Sciences

View full text Add to dashboard Cite

“…Alternatively, the combination of classifier outputs can be performed on an entire decision profile or the selected information to constrain a class decision. We refer to this alternative group of methods as class-indifferent methods [12]. In our work, the concept of the class-indifferent methods is slightly different from the ones aforementioned in [1], [12] and [23] .…”

Section: Modelling Classifier Outputs and Combination Methodsmentioning

confidence: 99%

“…The design of a method for combining classifier decisions is a challenging task in constructing an effective ensemble and various methods have been developed in the past decades. Kuncheva in [12] roughly characterizes combination methods, based on the forms of classifier outputs, into two categories. In the first category, the combination of decisions is performed on single classes, such as majority voting and Bayesian probability, which have extensively been examined in the ensemble literature [9], [11] and [23].…”

Section: Introductionmentioning

confidence: 99%

“…This method includes meta-learning − stacking where combination functions are learnt from numeric values of classes [10] and [17], linear sum and order statistics [11], [21] and [24] . An alternative group of methods called classindifferent methods makes use of as much information as possible obtained from sets of classes in calculating the support for each class [12].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Combining Classifiers through Triplet-Based Belief Functions

Shen

et al.

Machine Learning and Knowledge Discovery in Databases

View full text Add to dashboard Cite

Abstract. Classifier outputs in the form of continuous values have often been combined using linear sum or stacking, but little is generally known about evidential reasoning methods for combining truncated lists of ordered decisions. In this paper we introduce a novel class-indifferent method for combining such a kind of classifier decisions. Specifically we model each output given by classifiers on new instances as a list of ranked decisions that is divided into 2 subsets of decisions, which are represented by triplet-based belief functions and then are combined using Dempster's rule of combination. We present a formalism for triplet-based belief functions and establish a range of general formulae for combining these beliefs in order to arrive at a consensus decision. In addition we carry out a comparative analysis with an alternative representation dichotomous belief functions on the UCI benchmark data. We also compare our combination method with the popular methods of stacking, boosting, linear sum and majority voting over the same benchmark data to demonstrate the advantage of our approach.

show abstract

DESIGNING AN IF–THEN RULES‐BASED ENSEMBLE OF HETEROGENEOUS BANKRUPTCY CLASSIFIERS: A GENETIC ALGORITHM APPROACH

Davalos

Leng

Feroz

et al. 2014

Intell. Sys. Acc. Fin. Mgmt.

View full text Add to dashboard Cite

SUMMARY This paper proposes a framework for an ensemble bankruptcy classifier that uses if–then rules to combine the outputs from a heterogeneous set of classifiers. A genetic algorithm (GA) induces the rules using an asymmetric, cost‐sensitive fitness function that includes accuracy and misclassification costs. The GA‐based ensemble classifier outperforms individual classifiers and ensemble classifiers generated by other methods. The results of the classifier are in the form of if–then rules. We apply the approach to a balanced dataset and an imbalanced dataset. Both are composed of firms subject to financial distress and cited in the US Securities and Exchange Commission's Accounting and Auditing Enforcement Releases. Copyright © 2014 John Wiley & Sons, Ltd.

show abstract

Combining Classifiers: Soft Computing Solutions

Cited by 94 publications

References 58 publications

Eigenclassifiers for combining correlated classifiers

Eigenclassifiers for combining correlated classifiers

Combining Classifiers through Triplet-Based Belief Functions

DESIGNING AN IF–THEN RULES‐BASED ENSEMBLE OF HETEROGENEOUS BANKRUPTCY CLASSIFIERS: A GENETIC ALGORITHM APPROACH

Contact Info

Product

Resources

About