This is an early access version, the complete PDF, HTML, and XML versions will be available soon.
Open AccessArticle
A Semi-Supervised Active Learning Method for Structured Data Enhancement with Small Samples
by
Fangling Leng
Fangling Leng
Dr. Fangling Leng is a lecturer at Northeastern University. She was admitted to Northeastern in 1997 [...]
Dr. Fangling Leng is a lecturer at Northeastern University. She was admitted to Northeastern University in 1997 to study computer science and technology and graduated in 2001 with a bachelor's degree. In the same year, she was admitted to Northeastern University Graduate School to study computer application technology, engaged in data warehouse research, and graduated in 2004 with a master's degree. In the same year, she was admitted to Northeastern University Graduate School to study for a doctorate in computer software and theory, mainly engaged in data warehouse and OLAP research, and received a doctorate in 2008.
*,
Fan Li
Fan Li ,
Wei Lv
Wei Lv ,
Yubin Bao
Yubin Bao
Dr. Yubin Bao graduated from the Department of Computer Science at Central South University with a a [...]
Dr. Yubin Bao graduated from the Department of Computer Science at Central South University with a bachelor's degree in 1990 and received a master's degree and a doctorate from Northeastern University in 1993 and 2003, respectively. Dr. Bao’s main research interests are business intelligence, data mining, big data analysis, image processing, and graph data management.
,
Xiaofeng Liu
Xiaofeng Liu ,
Tiancheng Zhang
Tiancheng Zhang
Dr. Tiancheng Zhang is a professor and doctoral supervisor at Northeastern University, a
senior of [...]
Dr. Tiancheng Zhang is a professor and doctoral supervisor at Northeastern University, a
senior member of the China Computer Society, and an executive member of the
China Information System Committee. He received a Ph.D. degree in computer software and theory from Northeastern University (NEU), China. From May 2012 to April 2013, he
was a visiting scholar at the University of New South Wales, Australia. His main research interests include artificial intelligence, smart education, big data analysis and mining, and spatiotemporal data management technology. He is a senior member of the China Computer Federation (CCF).
and
Ge Yu
Ge Yu
Dr. Ge Yu received a PhD degree in computer science from the Kyushu University of Japan in 1996. He [...]
Dr. Ge Yu received a PhD degree in computer science from the Kyushu University of Japan in 1996. He is currently a professor at the Northeastern University of China. His research interests include distributed database systems, data science and big data, blockchain, and applications. He is a fellow of the China Computer Federation (CCF), a senior member of IEEE, and a senior member of ACM.
School of Computer Science and Engineering, Northeastern University, Shenyang 110169, China
*
Author to whom correspondence should be addressed.
Mathematics 2024, 12(17), 2634; https://doi.org/10.3390/math12172634 (registering DOI)
Submission received: 30 July 2024
/
Revised: 19 August 2024
/
Accepted: 22 August 2024
/
Published: 24 August 2024
Abstract
In order to solve the problems of the small capacity of structured data and uneven distribution among classes in machine learning tasks, a supervised generation method for structured data called WAGAN and a cyclic sampling method named SACS (Semi-supervised and Active-learning Cyclic Sampling), based on semi-supervised active learning, are proposed. The loss function and neural network structure are optimized, and the quantity and quality of the small sample set are enhanced. To enhance the reliability of generating pseudo-labels, a Semi-supervised Active learning Framework (SAF) is designed. This framework redistributes class labels to samples, which not only enhances the reliability of generated samples but also reduces the influence of noise and uncertainty on the generation of false labels. To mine the diversity information of generated samples, an uncertain sampling strategy based on spatial overlap is designed. This strategy incorporates the idea of spatial overlap and uses global and local sampling methods to calculate the information content of generated samples. Experimental results show that the proposed method performs better than other data enhancement methods on three different datasets. Compared to the original data, the average value of the classification model is improved by 11.5%, 16.1%, and 19.6% relative to compared methods.
Share and Cite
MDPI and ACS Style
Leng, F.; Li, F.; Lv, W.; Bao, Y.; Liu, X.; Zhang, T.; Yu, G.
A Semi-Supervised Active Learning Method for Structured Data Enhancement with Small Samples. Mathematics 2024, 12, 2634.
https://doi.org/10.3390/math12172634
AMA Style
Leng F, Li F, Lv W, Bao Y, Liu X, Zhang T, Yu G.
A Semi-Supervised Active Learning Method for Structured Data Enhancement with Small Samples. Mathematics. 2024; 12(17):2634.
https://doi.org/10.3390/math12172634
Chicago/Turabian Style
Leng, Fangling, Fan Li, Wei Lv, Yubin Bao, Xiaofeng Liu, Tiancheng Zhang, and Ge Yu.
2024. "A Semi-Supervised Active Learning Method for Structured Data Enhancement with Small Samples" Mathematics 12, no. 17: 2634.
https://doi.org/10.3390/math12172634
Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details
here.
Article Metrics
Article Access Statistics
For more information on the journal statistics, click
here.
Multiple requests from the same IP address are counted as one view.