An Ensemble Fuzziness-Based Online Sequential Learning Approach and Its Application

Cao, Weipeng; Li, Shengdong; Huang, Cheng-Chao; Wu, Yuhao; Wang, Qiang; Li, Da-Chuan; Liu, Ye

doi:10.1007/978-3-030-82136-4_21

Cited by 5 publications

(3 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We performed two experiments with the Data Analysis Unit. Although there are better techniques to analyze the data gathered as demonstrated in Cao et al (2018Cao et al ( , 2021a, we choose to use the K-means and the HDBSCAN clustering techniques for validation purposes. The performance of both techniques was not the objective of this paper.…”

Section: Prototype Componentsmentioning

confidence: 99%

Sapientia: a Smart Campus model to promote device and application flexibility

et al. 2022

View full text Add to dashboard Cite

Section: Prototype Componentsmentioning

confidence: 99%

Sapientia: a Smart Campus model to promote device and application flexibility

et al. 2022

View full text Add to dashboard Cite

“…Recently, fuzziness based SSL research [32]- [35] has also gained attention for SSL and several methods have been introduced [32]- [35]. In [32], authors proposed the new mechanism of data filtering from fuzziness prospective for classification problem on online sequential extreme learning machine (FOS-ELM).…”

Section: Related Workmentioning

confidence: 99%

“…In [34] a similar approach is proposed but the key difference is that this work [33] additionally handles the class imbalance problem using synthetic minority over-sampling technique, and used for liver diseases detection. In [35] authors proposed a novel method of filtering the data using ensemble fuzziness-based online sequential learning approach to update the single model on terminal. It consists of two module; one is a server module and other is a terminal module.…”

Section: Related Workmentioning

confidence: 99%

Binary-Classifiers-Enabled Filters for Semi-Supervised Learning

et al. 2021

View full text Add to dashboard Cite

A typical semi-supervised learning-based scheme is based on training a single model for labeled data. For unlabeled data, it uses the pseudo-labeling method to obtain labels. However, the samples during pseudolabeling are often filtered using a probability threshold, which suffers from the challenge of effective threshold selection. In the case of a high probability threshold, correct samples may not be labeled, and in the case of a low threshold, samples can be wrongly labeled. This threshold issue degrades the overall performance of the model. This paper addresses this vital issue by proposing a novel approach of SSL named Binary-Classifiers-Enabled Filters for Semi-Supervised Learning (BSSL) for labeling the unlabeled data by using binary classifiers as data filters. That is, we train binary classifiers dedicated to each class. After training, we propose three methods for labeling the unlabeled data; cascading, non-cascading, and rankbased binary classifiers. Our extensive experiment shows rank-based binary classifiers are the best choice for labeling the data. Our approach eliminates threshold selection to improve the performance of the model. Comprehensive experiments are performed to demonstrate the effectiveness of our approach on a variety of domains, including image classification, text classification and audio classification, datasets including MNIST, fashion-MNIST, EuroSat, ESC10, Free Spoken Digit dataset, Audio Emotion recognition, reuter and mice protein dataset. Rank based binary classifiers (BSSL) approach achieves absolute performance of atleast 10% and 5% over supervised learning(SL) and SSL, respectively on audio datasets in different number of sample cases except RAVDESS dataset. Moreover, BSSL shows tremendous performance on image datasets specifically when number of samples is very small. Overall, BSSL outperformed the purely supervised learning approach and SSL pseudo-labeling approaches in different number of samples cases.

show abstract

Sparse Dense Transformer Network for Video Action Recognition

Zhang

et al. 2022

Knowledge Science, Engineering and Management

View full text Add to dashboard Cite

An Ensemble Fuzziness-Based Online Sequential Learning Approach and Its Application

Cited by 5 publications

References 25 publications

Sapientia: a Smart Campus model to promote device and application flexibility

Sapientia: a Smart Campus model to promote device and application flexibility

Binary-Classifiers-Enabled Filters for Semi-Supervised Learning

Sparse Dense Transformer Network for Video Action Recognition

Contact Info

Product

Resources

About