Classifier ensembles for image identification using multi-objective Pareto features

Albukhanajer, Wissam A.; Jin, Yaochu; Briffa, Johann A.

doi:10.1016/j.neucom.2017.01.067

Cited by 39 publications

(16 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In Multi-Train, classifier diversity is gleaned by simultaneously manipulating data, manipulating input attributes, using various machine learning algorithms, and various models. Heterogeneous ensembles have proved to be more effective in achieving diversity [23,26,27], which is also empirically confirmed by statistically better results than those of the tri-training algorithm on the 12 datasets used in this work. Second, the proposed method predicts a probability of a data having a particular label, which can then be used to select the most confidently predicted unlabeled data to be added to labeled data.…”

Section: Introductionsupporting

confidence: 71%

See 1 more Smart Citation

Multi-train: A semi-supervised heterogeneous ensemble classifier

Jin

2017

Neurocomputing

Self Cite

View full text Add to dashboard Cite

Section: Introductionsupporting

confidence: 71%

“…To benefit from the improved accuracy of ensemble learning [20,21,22,23], techniques that combine SSL with ensembles have recently attracted much interest. For example, Shao and Tian [24] proposed a selective SSL ensemble learning method based on the distance to the model.…”

Section: Introductionmentioning

confidence: 99%

Multi-train: A semi-supervised heterogeneous ensemble classifier

Jin

2017

Neurocomputing

Self Cite

View full text Add to dashboard Cite

“…1. Heterogeneous ensembles that use both Input Input (1) Model (1) different types of models and different input features created by feature selection and feature extraction have shown better performance than those using single type of inputs [60], [61], partly due to the inherent better ability to create diversity [36]. Therefore, this work also employs different features and different types of models for promoting ensemble diversity without impairing accuracy, hoping to enhance the ability of the ensemble surrogate in uncertainty estimation.…”

Section: Heterogeneous Ensemble Generationmentioning

confidence: 99%

“…In machine learning, it has been theoretically proved that when a right balance between diversity and accuracy is met, an ensemble can provide more accurate predictions than any of its member alone [33]. Ensembles are often categorized into homogeneous ensembles (consisting of the same type of models typically generated by manipulating the data, like random sampling [34]) and heterogeneous ensembles (composed of different types of models [35] or different input features [36]).…”

Section: Introductionmentioning

confidence: 99%

Heterogeneous Ensemble-Based Infill Criterion for Evolutionary Multiobjective Optimization of Expensive Problems

Guo

Jin

Ding

et al. 2019

IEEE Trans. Cybern.

Self Cite

177

View full text Add to dashboard Cite

Gaussian processes (GPs) are the most popular model used in surrogate-assisted evolutionary optimization of computationally expensive problems, mainly because GPs are able to measure the uncertainty of the estimated fitness values, based on which certain infill sampling criteria can be used to guide the search and update the surrogate model. However, the computation time for constructing GPs may become excessively long when the number of training samples increases, which makes it inappropriate to use them as surrogates in evolutionary optimization. To address this issue, this paper proposes to use ensembles as surrogates and infill criteria for model management in evolutionary optimization. A heterogeneous ensemble consisting of a least square support vector machine and two radial basis function networks is constructed to enhance the reliability of ensembles for uncertainty estimation. In addition to the original decision variables, a selected subset of the decision variables and a set of transformed variables are used as inputs of the heterogeneous ensemble to further promote the diversity of the ensemble. The proposed heterogeneous ensemble is compared with a GP and a homogeneous ensemble for infill sampling criteria in evolutionary multiobjective optimization. Experimental results demonstrate that the heterogeneous ensemble is competitive in performance compared with GPs and much more scalable in computational complexity to the increase in search dimension.

show abstract

“…The definition of the best technique for a particular classification application is a challenging problem and has been addressed by several authors [15][16][17][18][19][20][21]. This problem has been treated as a meta-learning problem [17,22], automatic selection of machine learning (auto-ML) [16,18,20] or an optimization problem [15,19,21]. In [18], for instance, the authors used an auto-ML technique to define the best classification algorithm, along with the best set of parameters, for a specific classification problem.…”

Section: State-of-the-art: Optimization Techniques For Classifier Ensmentioning

confidence: 99%

Hybrid Metaheuristics to the Automatic Selection of Features and Members of Classifier Ensembles

2018

View full text Add to dashboard Cite

Metaheuristic algorithms have been applied to a wide range of global optimization problems. Basically, these techniques can be applied to problems in which a good solution must be found, providing imperfect or incomplete knowledge about the optimal solution. However, the concept of combining metaheuristics in an efficient way has emerged recently, in a field called hybridization of metaheuristics or, simply, hybrid metaheuristics. As a result of this, hybrid metaheuristics can be successfully applied in different optimization problems. In this paper, two hybrid metaheuristics, MAMH (Multiagent Metaheuristic Hybridization) and MAGMA (Multiagent Metaheuristic Architecture), are adapted to be applied in the automatic design of ensemble systems, in both mono- and multi-objective versions. To validate the feasibility of these hybrid techniques, we conducted an empirical investigation, performing a comparative analysis between them and traditional metaheuristics as well as existing existing ensemble generation methods. Our findings demonstrate a competitive performance of both techniques, in which a hybrid technique provided the lowest error rate for most of the analyzed objective functions.

show abstract

Classifier ensembles for image identification using multi-objective Pareto features

Cited by 39 publications

References 45 publications

Multi-train: A semi-supervised heterogeneous ensemble classifier

Multi-train: A semi-supervised heterogeneous ensemble classifier

Heterogeneous Ensemble-Based Infill Criterion for Evolutionary Multiobjective Optimization of Expensive Problems

Hybrid Metaheuristics to the Automatic Selection of Features and Members of Classifier Ensembles

Contact Info

Product

Resources

About