Fusion of instance selection methods in regression tasks

Arnaiz‐González, Álvar; Blachnik, Marcin; Kordos, Mirosław; García‐Osorio, César

doi:10.1016/j.inffus.2015.12.002

Cited by 36 publications

(16 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Recent methods in the field first solve other problems and then use classification, namely, regression problems as in [25][26][27][28][29][30], instance selection in data streams as in [31][32][33], and time series classification [34,35], or build ensembles of instance selection [36][37][38][39][40] and even create meta-learning systems, which automatically adjust a proper instance selection method to a given dataset as in [41,42].…”

Section: The Instance Selection Methodsmentioning

confidence: 99%

Instance Selection for Classifier Performance Estimation in Meta Learning

Blachnik

2017

Entropy

Self Cite

View full text Add to dashboard Cite

Abstract:Building an accurate prediction model is challenging and requires appropriate model selection. This process is very time consuming but can be accelerated with meta-learning-automatic model recommendation by estimating the performances of given prediction models without training them. Meta-learning utilizes metadata extracted from the dataset to effectively estimate the accuracy of the model in question. To achieve that goal, metadata descriptors must be gathered efficiently and must be informative to allow the precise estimation of prediction accuracy. In this paper, a new type of metadata descriptors is analyzed. These descriptors are based on the compression level obtained from the instance selection methods at the data-preprocessing stage. To verify their suitability, two types of experiments on real-world datasets have been conducted. In the first one, 11 instance selection methods were examined in order to validate the compression-accuracy relation for three classifiers: k-nearest neighbors (kNN), support vector machine (SVM), and random forest. From this analysis, two methods are recommended (instance-based learning type 2 (IB2), and edited nearest neighbor (ENN)) which are then compared with the state-of-the-art metaset descriptors. The obtained results confirm that the two suggested compression-based meta-features help to predict accuracy of the base model much more accurately than the state-of-the-art solution.

show abstract

Section: The Instance Selection Methodsmentioning

confidence: 99%

Instance Selection for Classifier Performance Estimation in Meta Learning

Blachnik

2017

Entropy

Self Cite

View full text Add to dashboard Cite

show abstract

“…Recently, the authors in [16] used modified versions of CNN and ENN for regression. Two strategies were presented; threshold-based (T-CNN, T-ENN), and discretization-based (D-CNN, D-ENN).…”

Section: Related Workmentioning

confidence: 99%

“…There are a few algorithms for regression tasks. Among them, the state-of-the-art work was introduced in [16]. However, this algorithm again focuses on maintaining an absolute error, instead of a relative error.…”

Section: Related Workmentioning

confidence: 99%

“…Since the proposed algorithm falls into the category of sample selection for regression tasks, two recent algorithms D-ENN, and D-CNN [16] were selected for comparison. Along with the above mentioned algorithms, Arnaiz-González et al, also proposed threshold based algorithms i.e., T-ENN, T-CNN, and their ensembles.…”

Section: Experimental Evaluationmentioning

confidence: 99%

“…Additionally, since the whole signal had to be reconstructed in each iteration, the time complexity was also increased. More recently, Arnize et al proposed Discretization-based Condensed Nearest Neighbor (D-CNN) and Discretization-based Edited Nearest Neighbor (D-ENN) as modified versions of Condensed Nearest Neighbor (CNN) and Edited Nearest Neighbor (ENN) for regression tasks [16]. Both of the algorithms ensure a balanced selection of samples throughout the sample space using equal-width binning of the uni-dimensional output signal.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Stimuli-Magnitude-Adaptive Sample Selection for Data-Driven Haptic Modeling

Abdulali

Hassan

Jeon

2016

Entropy

View full text Add to dashboard Cite

Data-driven haptic modeling is an emerging technique where contact dynamics are simulated and interpolated based on a generic input-output matching model identified by data sensed from interaction with target physical objects. In data-driven modeling, selecting representative samples from a large set of data in a way that they can efficiently and accurately describe the whole dataset has been a long standing problem. This paper presents a new algorithm for the sample selection where the variances of output are observed for selecting representative input-output samples in order to ensure the quality of output prediction. The main idea is that representative pairs of input-output are chosen so that the ratio of the standard deviation to the mean of the corresponding output group does not exceed an application-dependent threshold. This output-and standard deviation-based sample selection is very effective in applications where the variance or relative error of the output should be kept within a certain threshold. This threshold is used for partitioning the input space using Binary Space Partitioning-tree (BSP-tree) and k-means algorithms. We apply the new approach to data-driven haptic modeling scenario where the relative error of the output prediction result should be less than a perceptual threshold. For evaluation, the proposed algorithm is compared to two state-of-the-art sample selection algorithms for regression tasks. Four kinds of haptic related behavior-force datasets are tested. The results showed that the proposed algorithm outperformed the others in terms of output-approximation quality and computational complexity.

show abstract

Haptic Software Design

Abdulali

Jeon

2022

Springer Series on Touch and Haptic Systems

View full text Add to dashboard Cite

This chapter reviews design concepts of haptic modeling and rendering software. The main focus lies in realistic kinesthetic and tactile haptic models for virtual and augmented reality based on the data collected from physical objects. We consider both data-driven algorithms providing a black-box action-response mapping and measurement-based approaches identifying parameters of physics-based models. To show the research landscape and highlight ongoing research challenges, we introduce a series of state-of-the-art methods including data-driven models with deterministic and stochastic responses, physics-based simulation using optimization-based FEM solver, and hybrid approaches of combining the concepts of both data-driven and physics-based methods. These examples also cover a wide range of haptic properties, i.e., modeling and rendering of elasticity and plasticity, tool deformation, and haptic textures.

show abstract

Fusion of instance selection methods in regression tasks

Cited by 36 publications

References 29 publications

Instance Selection for Classifier Performance Estimation in Meta Learning

Instance Selection for Classifier Performance Estimation in Meta Learning

Stimuli-Magnitude-Adaptive Sample Selection for Data-Driven Haptic Modeling

Haptic Software Design

Contact Info

Product

Resources

About