14,459 Hits in 4.7 sec

Handling uncertainties in SVM classification [article]

Emilie Niaf, Carole Lartizien
2011 arXiv   pre-print
Based on empirical evidence our method outperforms regular SVM in terms of probability predictions and classification performances.  ...  This paper addresses the pattern classification problem arising when available target data include some uncertainty information.  ...  classification), • real values: l i = p i ∈ [0, 1] for i = n + 1 . . . m (in regression). p i , associated to point x i allows to consider uncertainties about point x i 's class.  ... 
arXiv:1106.3397v1 fatcat:5kuhiifq6jf7dc56bt6fq4nv4u

Handling uncertainties in SVM classification

Emilie Niaf, Remi Flamary, Carole Lartizien, Stephane Canu
2011 2011 IEEE Statistical Signal Processing Workshop (SSP)  
Based on empirical evidence our method outperforms regular SVM in terms of probability predictions and classification performances.  ...  This paper addresses the pattern classification problem arising when available target data include some uncertainty information.  ...  classification), • real values: l i = p i ∈ [0, 1] for i = n + 1 . . . m (in regression). p i , associated to point x i allows to consider uncertainties about point x i 's class.  ... 
doi:10.1109/ssp.2011.5967814 fatcat:o5qpq24m6baptpf5mfosyt6fwe

Ensembles of Deep Learning Framework for Stomach Abnormalities Classification

Talha Saeed, Chu Kiong Loo, Muhammad Shahreeza Safiruz Kassim
2022 Computers Materials & Continua  
To play an essential role in real-time medical diagnosis, CNN-based models need to be accurate and interpretable, while the uncertainty must be handled.  ...  At last, the features are fused, and uncertainties are handled by selecting entropybased features.  ...  Also, they did not focus on uncertainty handling. We are the first to integrate them into GIT abnormality classification.  ... 
doi:10.32604/cmc.2022.019076 fatcat:cx2qs3rzgrhrlobc47tbd4ovhu

A Type-2 Fuzzy Logic Ensemble SVM Classifier based on Feature Weighting

Weiping Huang, Ziyang Wang, Qinghua Chi, Jun Liang
2016 International Journal of Signal Processing, Image Processing and Pattern Recognition  
Yet the mapping relationships given out is already qualified to verify type-2 fuzzy logic's capability to handle uncertainty in the ensemble SVM.  ...  Considering that the SVM classification accuracy is easily affected by outliers and noise in training data, thus there exists uncertainty in relationships between each single SVM's classification accuracy  ...  Interval type-2 fuzzy logic is adopted in the ensemble classifier of this paper. Type-2 fuzzy logic performs much better than type-1 in handling uncertainty.  ... 
doi:10.14257/ijsip.2016.9.2.28 fatcat:3jrs27a6tjcy5nnllteshqkupu

Multilabel Classification using Bayesian Compressed Sensing

Ashish Kapoor, Raajay Viswanathan, Prateek Jain
2012 Neural Information Processing Systems  
The two key benefits of the model are that a) it can naturally handle datasets that have missing labels and b) it can also measure uncertainty in prediction.  ...  In this paper, we present a Bayesian framework for multilabel classification using compressed sensing.  ...  Non-probabilistic classification schemes, such as SVMs, can handle traditional active learning by first establishing the confidence in the estimate of each label by using the distance from the classification  ... 
dblp:conf/nips/KapoorVJ12 fatcat:rne7g3z75bddtpts5mc7fs3hhq

Robust Support Vector Machines with Low Test Time

Yahya Forghani, Hadi Sadoghi Yazdi
2014 Computational intelligence  
In this article, we show that the primal RoSVM is equivalent to an SOCP and name it accurate primal RoSVM. The optimal weight vector of this model is not sparse necessarily.  ...  Second, we show that some parts of the optimal decision function can be computed in the training phase instead of the test phase. This can decrease the test time further.  ...  Recent studies have shown that robust SVM (RoSVM), which explicitly handles the data uncertainties, can improve the performance of the traditional SVM (Bhattacharyya, Grate, et al. 2004a; Bhattacharyya  ... 
doi:10.1111/coin.12039 fatcat:poi4h43urbgghilspvshfiaori

Uncertainty Handling in Model Selection for Support Vector Machines [chapter]

Tobias Glasmachers, Christian Igel
2008 Lecture Notes in Computer Science  
In the next section we introduce the CMA-ES and the noise-handling technique. In Section 3 we briefly review SVMs for binary classification. Then we motivate our model selection objective function.  ...  Our results show that this search strategy avoids premature convergence and results in improved classification accuracy compared to strategies without uncertainty handling. mechanism has been proposed  ...  We applied the CMA-ES with and without uncertainty handling mechanism to the problem of model selection for SVMs.  ... 
doi:10.1007/978-3-540-87700-4_19 fatcat:juitrdjpyncrljc7mnxhxtzm6m

Robust Cost Sensitive Support Vector Machine

Shuichi Katsumata, Akiko Takeda
2015 International Conference on Artificial Intelligence and Statistics  
In general, robust classifications are used to create a classifier robust to data by taking into account the uncertainty of the data.  ...  In such cases, it is common for the uncertainty of the data to be characterized by  ...  In contrast, in our robust SVM model we assign larger uncertainty sets on the minority class and assign smaller uncertainty sets on the majority class.  ... 
dblp:conf/aistats/KatsumataT15 fatcat:m2snelqqrbeuxigsf5o3uvkbhq

Survey on uncertainty support vector machine and its application in fault diagnosis

Yi-Bo Li, Ye Li
2010 2010 3rd International Conference on Computer Science and Information Technology  
In this paper, first, the combination of basic uncertainty mathematics theory and support vector machine (SVM) and its application in fault diagnosis are introduced in detail.  ...  Because of the superiority on processing the uncertain information and fuzzy information, the uncertainty mathematical theory has been widely applied in fault diagnosis of complex system.  ...  THE APPLICATION OF UNCERTAINTY SVM Uncertainty SVM is a new research field about commerce and data.  ... 
doi:10.1109/iccsit.2010.5563619 fatcat:oifn4mdbfjec3l5pgszdbrsiby

Performance Study Of Uncertainty Based Feature Selection Method On Detection Of Chronic Kidney Disease With SVM Classification

Lailly Syifa'ul Qolby, Joko Lianto Buliali, Ahmad Saikhu
2021 IPTEK: The Journal for Technology and Science  
Uncertainty (MSU) Method In the third trial, SVM classification uses the Multivariate Symmetrical Uncertainty (MSU) feature selection method.  ...  Symmetrical Uncertainty (SU) and Multivariate Symmetrical Uncertainty (MSU) are feature selection methods that give a different classification result using SVM. 3.1 Dataset The dataset used in this  ... 
doi:10.12962/j20882033.v32i2.10483 fatcat:zmpmqkzhejebdpp6iqhuh2qhqy

Application of a New Loss Function-Based Support Vector Machine Algorithm in Quality Control of Measurement Observation Data

Youping Wu, Guoqiang Tao
2022 Mathematical Problems in Engineering  
The results show that the SVM algorithm with the new loss function has better accuracy in processing the observed data.  ...  The loss function of traditional SVM methods is that the uncertainty distribution of the data cannot be obtained.  ...  Acknowledgments is work was supported by Science and Technology Research Project of Jiangxi Provincial Department of Education: Application of multiplicative and additive mixed noise model in SAR (or Insar  ... 
doi:10.1155/2022/7266719 doaj:5cdc9610facb4792925b3d9900965d16 fatcat:w5yutfgdyja2jg3jptddatcrfm

Linear Maximum Margin Classifier for Learning from Uncertain Data

Christos Tzelepis, Vasileios Mezaris, Ioannis Patras
2017 IEEE Transactions on Pattern Analysis and Machine Intelligence  
In this paper, we propose a maximum margin classifier that deals with uncertainty in data input.  ...  We address the classification problem and define a cost function that is the expected value of the classical SVM cost when data samples are drawn from the multi-dimensional Gaussian distributions that  ...  We show that our method outperforms the linear SVM and other SVM variants that handle uncertainty isotropically.  ... 
doi:10.1109/tpami.2017.2772235 pmid:29990153 fatcat:6snjmdesgnacrioo5lqpi3vx6i

An Uncertainty Framework for Classification [article]

Loo-Nin Teow, Kia-Fock Loe
2013 arXiv   pre-print
In the probabilistic framework, we obtain classifiers that optimize the cross-entropy function. In the possibilistic framework, we obtain classifiers that maximize the interclass margin.  ...  We define a generalized likelihood function based on uncertainty measures and show that maximizing such a likelihood function for different measures induces different types of classifiers.  ...  INTRODUCTION Uncertainty is a natural and unavoidable part of pattern classification in real-world domains.  ... 
arXiv:1301.3896v1 fatcat:gt4wgk3pbzgmjiqmmoknt5bv6m

Robust Formulations for Handling Uncertainty in Kernel Matrices

Sahely Bhadra, Sourangshu Bhattacharya, Chiranjib Bhattacharyya, Aharon Ben-Tal
2010 International Conference on Machine Learning  
We study the problem of uncertainty in the entries of the Kernel matrix, arising in SVM formulation.  ...  The problem of uncertainty in kernel matrix is motivated from the real world problem of classifying proteins when the structures are provided with some uncertainty.  ...  SVM, are unable to handle uncertainty, compared to the proposed robust classifiers. RSVM QP performs comparably with SVM since the assumption of β being rank 1 does not hold for the current dataset.  ... 
dblp:conf/icml/BhadraBBB10 fatcat:vqwm7ez34bg3hcvllhdq3dufni

Multi-class active learning for image classification

A.J. Joshi, F. Porikli, N. Papanikolopoulos
2009 2009 IEEE Conference on Computer Vision and Pattern Recognition  
Specifically, we propose an uncertainty measure that generalizes margin-based uncertainty to the multi-class case and is easy to compute, so that active learning can handle a large number of classes and  ...  One of the principal bottlenecks in applying learning techniques to classification problems is the large amount of labeled training data required.  ...  Acknowledgment This material is based upon work supported in part by the U.S. Army Research Laboratory and the U.S.  ... 
doi:10.1109/cvprw.2009.5206627 fatcat:hgd6pkiodjeqjiokq5d2r5vaky
« Previous Showing results 1 — 15 out of 14,459 results