GPU-Accelerated CatBoost-Forest for Hyperspectral Image Classification Via Parallelized mRMR Ensemble Subspace Feature Selection

Samat, Alim; Li, Erzhu; Du, Peijun; Liu, Sicong; Xia, Junshi

doi:10.1109/jstars.2021.3063507

Cited by 38 publications

(18 citation statements)

References 75 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For the multi‐category problem of converter valve state assessment, 1, 2, 3, 4 respectively represent normal, attention, abnormal and severe states. The main steps of the CatBoost algorithm are as follows [25]: For each sample p i in the training set X , CatBoost will utilize all the samples except it to train and obtain the model M i . …”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

A data‐driven approach to state assessment of the converter valve based on oversampling and Shapley additive explanations

Mei

Zheng

et al. 2022

IET Generation Trans & Dist

View full text Add to dashboard Cite

The utilization of data-driven artificial intelligence technology for ultra-high voltage converter valve state assessment can improve efficiency and accuracy, but there are still problems such as imbalance of raw data and poor model interpretability. This paper aims to propose a new type of ultra-high voltage converter valve state assessment method to solve these problems. An oversampling method that considers characteristic boundary information is proposed, which can retain and enhance the boundary information of minority samples. The balanced data are imported into categorical boosting to realize offline training and online evaluation of converter valve state assessment. An interpretability analysis framework of the converter valve state assessment model based on the Shapley additive explanations is proposed, which can explain the operation logic of the model and find the key factors that affect the assessment results. The calculation example shows that the evaluation accuracy of this method can reach up to 98.5%, and the running speed is faster than the conventional method. The results can provide guidance for the maintenance and repair of UHV converter valves.

show abstract

Section: Methodsmentioning

confidence: 99%

“…For the multi-category problem of converter valve state assessment, 1, 2, 3, 4 respectively represent normal, attention, abnormal and severe states. The main steps of the CatBoost algorithm are as follows [25]:…”

Section: Boosting and Catboost Classifiersmentioning

confidence: 99%

A data‐driven approach to state assessment of the converter valve based on oversampling and Shapley additive explanations

Mei

Zheng

et al. 2022

IET Generation Trans & Dist

View full text Add to dashboard Cite

show abstract

“…CatBoost is an effective method for converting categorical data into numeric data and preventing overfitting. Categorical data are mainly preprocessed according to the following three steps (Hancock & Khoshgoftaar, 2020;Samat et al, 2021;Zhang et al, 2021).…”

Section: Catboost Algorithmmentioning

confidence: 99%

Rapid determination of water content in potato tubers based on hyperspectral images and machine learning algorithms

Zou

Chen

et al. 2022

Food Sci. Technol

View full text Add to dashboard Cite

This study investigated the hyperspectral reflectance response of time series generated during oven drying to changes in the moisture content of potato tubers. Seventeen preprocessing methods were used to eliminate the influence of spectral noise on the spectral characteristic curve. Algorithms such as CatBoost, LightGBM, and XGBoost are used to obtain the first 40 effective characteristic spectra of hyperspectral images, which reduces the redundancy of data and improves the prediction accuracy. The water content prediction model of potato tubers was established by using the selected characteristic bands. The results showed that the combined model based on Lasso and XGBoost algorithm had the strongest prediction ability. The best model is MF-Lasso-XGBoost, which has R 2 value of 0.8908, Rmse of 0.0610, Mdae of 0.0389, and R 2 cv of 0.8448. This research can provide reference for the detection of potato moisture content and theoretical basis for the development of crop moisture detector.

show abstract

“…That is, a larger weight is given to the misclassified samples. Then a new base classifier is trained based on the adjusted sample distribution, and this step is repeated until the model converges [24], [25]. The most famous algorithm in boosting methods is Adaboost, which has many applications in hyperspectral image classification.…”

Section: Introductionmentioning

confidence: 99%

Dynamic Ensemble Learning With Multi-View Kernel Collaborative Subspace Clustering for Hyperspectral Image Classification

Lü

et al. 2022

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Recently, a series of collaborative representation (CR) methods have attracted much attention for hyperspectral images classification. In this paper, two CR-based dynamic ensemble selection (DES) methods using multi-view kernel collaborative subspace clustering (MVKCSC) and random subspace multi-view kernel collaborative subspace clustering (RSMVKCSC) are proposed. In order to combine spectral and spatial information to construct a region of competence (RoC), the multi-view learning strategy is used in the general DES method. Compared with traditional clustering methods, the MVC can more effectively utilize multi-feature information. Moreover, a new method of constructing the Laplacian matrix using kernel collaborative representation coefficients is proposed for clustering based on subspace clustering and CR theory. This method is called MVKCSC, which can obtain the clustering results by using kernel CR self-representation coefficients. In addition, to increase diversity of samples, the random subspace method (RSM) and MVKCSC are combined for RMVKCSC. Moreover, the algorithm can obtain better clustering results by constraining samples and features simultaneously. The effectiveness of the proposed methods is validated using three hyperspectral data sets with few samples. The experimental results show that both DES-MVKCSC and DES-RSMVKCSC outperform their single classifier counterparts. In particular, the proposed methods provide superior performance compared with the state-of-the-art DES methods.

show abstract

GPU-Accelerated CatBoost-Forest for Hyperspectral Image Classification Via Parallelized mRMR Ensemble Subspace Feature Selection

Cited by 38 publications

References 75 publications

A data‐driven approach to state assessment of the converter valve based on oversampling and Shapley additive explanations

A data‐driven approach to state assessment of the converter valve based on oversampling and Shapley additive explanations

Rapid determination of water content in potato tubers based on hyperspectral images and machine learning algorithms

Dynamic Ensemble Learning With Multi-View Kernel Collaborative Subspace Clustering for Hyperspectral Image Classification

Contact Info

Product

Resources

About