C-parameter version of robust bounded one-class support vector classification.

Sci Rep

College of Mathematics and Systems Science, Xinjiang University, Urumqi , 830046, China.

Published: January 2025

ν-one-class support vector classification (ν-OCSVC) has garnered significant attention for its remarkable performance in handling single-class classification and anomaly detection. Nonetheless, the model does not yield a unique decision boundary, and potentially compromises learning performance when the training data is contaminated by some outliers or mislabeled observations. This paper presents a novel C-parameter version of bounded one-class support vector classification (C-BOCSVC) to determine a unique decision boundary. The distance from the origin to decision boundary is the geometrical margin in the space [Formula: see text] (higher 1-dimension than feature space), and its maximization corresponds to the structural risk minimization (SRM) principle inscribed by an [Formula: see text]-norm regularization term both on the normal direction and bias of the decision boundary. To enhance the anti-noise and anti-outlier abilities of C-BOCSVC, the alternative robust version (C-RBOCSVC) is also developed, which incorporates the k-nearest neighbor relative density to assign varying weights to observations and mitigate the negative impact of outliers on the optimal decision boundary. The theoretical properties of the proposed method are successively derived, including the relationship between the solutions to the primal and dual problems, the connections between our C-BOCSVC and ν-OCSVC and the computational complexity. Experimental results over massive datasets demonstrate the feasibility and reliability of our C-BOCSVC, and highlight the superior performance of C-RBOCSVC compared to other state-of-the-art one-class classifiers when data is contaminated. The demo code of this work is publicly available at https://github.com/Zhangmath1122/C-RBOCSVC .

Download full-text PDF

Source
http://dx.doi.org/10.1038/s41598-025-85151-7DOI Listing

Publication Analysis

Top Keywords

decision boundary
20
support vector
12
vector classification
12
c-parameter version
8
bounded one-class
8
one-class support
8
unique decision
8
data contaminated
8
decision
5
boundary
5

Similar Publications

Prompt learning is a powerful technique that enables the transfer of Vision-Language Models (VLMs) like CLIP to downstream tasks. However, when the prompt-based methods are fine-tuned solely on base classes, they often struggle to generalize to novel classes lacking visual samples during training, especially in scenarios with limited training data. To address this challenge, we propose an innovative approach called Synth-CLIP that leverages synthetic data to enhance CLIP's generalization capability for base classes and the general capability for novel classes.

View Article and Find Full Text PDF

C-parameter version of robust bounded one-class support vector classification.

Sci Rep

January 2025

College of Mathematics and Systems Science, Xinjiang University, Urumqi , 830046, China.

ν-one-class support vector classification (ν-OCSVC) has garnered significant attention for its remarkable performance in handling single-class classification and anomaly detection. Nonetheless, the model does not yield a unique decision boundary, and potentially compromises learning performance when the training data is contaminated by some outliers or mislabeled observations. This paper presents a novel C-parameter version of bounded one-class support vector classification (C-BOCSVC) to determine a unique decision boundary.

View Article and Find Full Text PDF

Identifying landscape patterns conducive to pollutant transport control is of vitally importance for water quality protection. However, it remains unclear which landscape patterns can weaken the transport capacity of pollutants entering water bodies. To fill this gap, this study proposes a new framework.

View Article and Find Full Text PDF

Simulating fish autonomous swimming behaviours using deep reinforcement learning based on Kolmogorov-Arnold Networks.

Bioinspir Biomim

January 2025

Chongqing Jiaotong University, No. 66, Xuefu Avenue, Nanan District, Chongqing City, Chongqing, Chongqing, 400074, CHINA.

The study of fish swimming behaviours and locomotion mechanisms holds significant scientific and engineering value. With the rapid advancements in artificial intelligence, a new method combining deep reinforcement learning (DRL) with computational fluid dynamics (CFD) has emerged and been applied to simulate the autonomous behavior of higher organisms like fish. However, the scale of this cross-disciplinary method is directly affected by the efficiency of the DRL model.

View Article and Find Full Text PDF

As we assess the habitability of other worlds, we are limited by being able to only study terrestrial life adapted to terrestrial conditions. The environments found on Earth, though tremendously diverse, do not approach the multitude of potentially habitable environments beyond Earth, and so limited terrestrial adaptive capabilities tell us little about the fundamental biochemical boundaries of life. One approach to this problem is to use experimental laboratory evolution to adapt microbes to these novel environmental conditions.

View Article and Find Full Text PDF

Want AI Summaries of new PubMed Abstracts delivered to your In-box?

Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!