Dongying Bai

Learn More
How to deal with the newly added training samples, and utilize the result of the previous training effectively to get better classification result fast are the main tasks of incremental learning. A fast SVM incremental learning algorithm based on the central convex hulls algorithm is proposed in this paper. To utilize the result of the previous training and(More)
To reduce the computational cost of the incremental learning, a fast SVM incremental learning algorithm based on the convex hulls algorithm is proposed in this paper. The given algorithm is based on utilizing the result of the previous training effectively and retaining the most important samples for the incremental learning to reduce the computational(More)
Support vector machine (SVM) has been used in high resolution range profile (HRRP) classification for its good generalization ability for the pattern classification problem with high feature dimension and small training set. In order to perform multi-class classification, decision-tree-based SVM was studied, the structure and the classification performance(More)
Membrane vesicles (MVs) of Porphyromonas gingivalis are regarded as an offensive weapon of the bacterium, leading to tissue deterioration in periodontal disease. Therefore, isolation of highly purified MVs is indispensable to better understand the pathophysiological role of MVs in the progression of periodontitis. MVs are generally isolated by a(More)
Multiple kernel learning is a new research focus in the field of kernel machine learning in recent years. Localized multiple kernel learning is a promising strategy for combining multiple features or kernels in terms of their discriminative power for different local space. In this paper, we proposed a group based non-sparse localized multiple kernel(More)
Decision directed acyclic graph support vector machine (DDAGSVM) has been proposed to extend SVM from binary classification problems to multi-class classifications. But the generalization ability is subject to the structure of DDAG. To improve the classification accuracy, a novel separability measure is defined based on Karush-Kuhn-Tucher (KKT) condition,(More)
By choosing the most informative patterns that have the most possibility to become the support vectors in the training data by using the convex hulls algorithm, a fast training algorithm for SVM (QhullSVM) is given in this paper. Experimental results reveal that the given QhullSVM has better training performance comparing with the traditional training(More)
From the geometric point of view and by choosing the most informative patterns that have the most possibility to become the support vectors in the training data by using the convex hulls algorithm, a fast training algorithm for SVM is given in this paper. In this training algorithm for SVM, the convex hull vectors are chosen firstly, and the convex hull(More)
  • 1