Dictionary-based classifiers for exploiting feature sequence information and their application to hyperspectral remotely sensed data

Patro, Ram Narayan; Subudhi, Subhashree; Biswal, Pradyut Kumar; Dell’Acqua, Fabio

doi:10.1080/01431161.2019.1577580

Cited by 18 publications

(10 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Then, the performance of FWAdaBoost-SOFIES is compared with a number of classification approaches on the nine multi-class datasets for benchmark comparison. The following nine widely used classification approaches are used for comparison (the first three are zero-order FISs): 1) selforganizing fuzzy inference ensemble system (SOFEnsemble) [53]; 2) zero-order autonomous learning multiple-model classifier (ALMMo0) [54]; 3) eClass0 classifier [39]; 4) SVM with linear kernel; 5) KNN; 6) DT; 7) multilayer perceptron (MLP); 8) sequence classifier (SC) [58], and; 9) extreme learning machine (ELM) [59].…”

Section: Performance Demonstrationmentioning

confidence: 99%

See 1 more Smart Citation

Multiclass Fuzzily Weighted Adaptive-Boosting-Based Self-Organizing Fuzzy Inference Ensemble Systems for Classification

Gu¹,

Angelov²

2022

IEEE Trans. Fuzzy Syst.

View full text Add to dashboard Cite

Section: Performance Demonstrationmentioning

confidence: 99%

“…In this example, SOFEnsemble, ALMMo0, eClass0 and SC classifiers follow the recommended parameter settings given by [39], [53], [54], [58]. For SVM, KNN, MLP, DT and ELM classifiers, five different parameter settings are considered for each of them during the experiments and the best performances are reported.…”

Section: Performance Demonstrationmentioning

confidence: 99%

Multiclass Fuzzily Weighted Adaptive-Boosting-Based Self-Organizing Fuzzy Inference Ensemble Systems for Classification

Gu¹,

Angelov²

2022

IEEE Trans. Fuzzy Syst.

View full text Add to dashboard Cite

“…Various modifications of such tree-based methods are presented in [35,36]. Dictionary-based classifiers [37] are also more prevalent in remote sensing but are mostly dependent on the rules to build the dictionary and may not be applicable for all applications. Further, if multiple trees are incorporated, they can be treated as forestbased approaches.…”

Section: Classifier Selectionmentioning

confidence: 99%

“…Instead of direct distances, a feature reconstruction error driven through regularization is also helpful to nonlinearly learn the boundaries. The nearest regularized subspace (NRS) classifier is such an approach [41] and has been investigated in the literature [37] for hyperspectral applications. For highly unbalanced data, the synthetic minority oversampling technique (SMOTE) [42,43] has been verified for classification.…”

Section: Classifier Selectionmentioning

confidence: 99%

NLOS identification and mitigation in UWB positioning with bagging-based ensembled classifiers

Rayavarapu

Mahapatro

2021

Ann. Telecommun.

View full text Add to dashboard Cite

The identification and mitigation of nonline-of-sight (NLOS) paths play crucial roles in effectively localizing sensor nodes deployed in both indoor and urban outdoor environments. In this work, the NLOS identification and mitigation tasks are performed using the bagging-based ensembled classifier. The proposed method is compared to other stateof-the-art approaches, and promising results, such as higher classification accuracy, are obtained. The received signal waveform is used as raw data, and further statistical features are extracted from the channel impulse response (CIR). The classification performances are analyzed over varying feature subsets, and associated hyperparameters are validated on three datasets. The obtained results strongly suggest implementing the bagging classifier with a proper selection of features and hyperparameters.

show abstract

“…Several approaches of BS can be categorized into supervised (Patro, Subudhi, et al, 2019a) (Yang, Du, Su, & Sheng, 2010) semi-supervised as well as unsupervised (Patra et al 2015) techniques. Again, depending on the type of information used, the process can be spectral (Patro, Subudhi, et al, 2019b) or spatial. In the present literature, a clustering and ranking-centric unsupervised BS is proposed and similar articles are discussed further.…”

Section: Introductionmentioning

confidence: 99%

Spatial residual clustering and entropy based ranking for hyperspectral band selection

Raju

Varma²,

Lakshmi

2019

European Journal of Remote Sensing

View full text Add to dashboard Cite

Though the Hyper-spectral images (HSI) are associated with rich spectral information for discriminating the class-specific objects, the high dimensional data generates Hughes effect for additional processing. So, during pre-processing, band Selection (BS) is done to lower the dimension. The proposed unsupervised BS technique follows '3ʹ subsequent steps, such as 1) spatial residual (Gaussian) filtering, 2) spatial feature-based band clustering (using K-means) and 3) Entropy-based ranking. Initially, the bands are filtered and the self residual images are obtained. The filtered residual images are mapped to k clusters, where each cluster represents unique spatial information. Further, each cluster undergoes selection through the informationtheoretic (Entropy) approach. Also, from the selected optimal bands, SpatioSpectral information is extracted to appraise the performance using Support Vector Machine (SVM) classifiers with different state-of-art approaches. Performance measures like Average Accuracy (AA), Kappa (κ) and Overall Accuracy (OA) are contrasted and are presented in tables. The experiment shows propitious results as contrasted to other approaches, and suggests that pre clustering of the bands is informative as the adjacent bands are strongly correlated. Also, on comparing other approaches, the proposed one is computationally cheap and faster, which may best suit for online (in-camera) BS purposes.

show abstract

Dictionary-based classifiers for exploiting feature sequence information and their application to hyperspectral remotely sensed data

Cited by 18 publications

References 38 publications

Multiclass Fuzzily Weighted Adaptive-Boosting-Based Self-Organizing Fuzzy Inference Ensemble Systems for Classification

Multiclass Fuzzily Weighted Adaptive-Boosting-Based Self-Organizing Fuzzy Inference Ensemble Systems for Classification

NLOS identification and mitigation in UWB positioning with bagging-based ensembled classifiers

Spatial residual clustering and entropy based ranking for hyperspectral band selection

Contact Info

Product

Resources

About