Label distribution learning (LDL) as a soft-labeling paradigm is allowed to learn single or multi-labeled information distribution. Overwhelmingly, in the open world, the distribution of labels is ...usually disturbed by the noise (such as the man-made induction bias, shake of hardware devices), which in turn affects the decision of downstream tasks. To address this problem, we propose a novel LDL approach by using the three-way decisions theory to clear the amplified noise in this paper. First, we evaluate the confidence of each training sample and use a three-way decisions-based method to identify the trustworthy samples and the noisy samples. Second, we apply the sample correlation between the trustworthy samples and the noisy samples to correct the noisy labels. Finally, we re-weight every sample based on the learned confidences to train the robust LDL model. Experiments show that our approach has better performance in handling noisy data compared to existing algorithms.
•Presence of noise in a data set misguides classifiers when data set is resampled by SMOTE as more noise is generated.•The number of links in SMOTE is vaguely selected and same for every ...observation.•We propose a new noise detection method to be applied before SMOTE to prevent noise generation.•We also propose a new approach to select the number of links automatically in SMOTE.•Proposed SMOTEWB method outperforms SMOTE in linear and nonlinear classifiers in presence of noise.
Most of the classification methods assume that the numbers of class observations are balanced. In such cases, models are predicted by giving biased weight to the the class with more observations. Therefore, the classifiers ignore the class with smaller number of observations and the majority class makes biased predictions. There are some advised performance measures to be used in datasets, as well as recommended approaches to solve class imbalance problem. One of the most widely used methods is resampling method. In this study, the difficulties relevant to random oversampling (ROS) and synthetic minority oversampling technique (SMOTE), which are some of the oversampling methods, are discussed. This study aims to propose a combination of a new noise detection method and SMOTE to overcome those difficulties. Using the boosting procedure in ensemble algorithms, noise detection is possible with the proposed SMOTE with boosting (SMOTEWB) method, which makes use of this information to determine the appropriate number of neighbors for each observation within SMOTE algorithm.
We present ConSent, a novel context-based approach for the task of sentiment analysis. Our approach builds on techniques from the field of information retrieval to identify key terms indicative of ...the existence of sentiment. We model these terms and the contexts in which they appear and use them to generate features for supervised learning. The two major strengths of the proposed model are its robustness against noise and the easy addition of features from multiple sources to the feature set. Empirical evaluation over multiple real-world domains demonstrates the merit of our approach, compared to state-of the art methods both in noiseless and noisy text.
Currently, few samples and the inevitable noise poses a severe test on deep learning methods. To solve the above problems, a novel fault diagnosis network based on a refined prototype and correlation ...weighting Manhattan distance (RPCMN) is proposed. Specifically, a multiscale feature extraction (MSFE) module and a sparse non-local attention (SNLA) module are developed to comprehensively extract key classification information. Moreover, a prototype interactive refinement mechanism (PIRM) is established to refine the position of prototypes to make them more representative. A correlation weighting Manhattan distance (CWMD) is designed to accentuate the correlation between different prototypes. The superiority of our method is verified on two standard datasets and one vibration dataset in practical industrial applications. We found that the diagnosis accuracy is 99.12 % at the training set size of 20. Meanwhile, at different noise levels (−6 to 6 dB), the diagnostic accuracy is higher than 90 %.
Feature selection has aroused extensive attention and aims at selecting features that are highly relevant to classification from raw datasets to improve the performance of a learning model. Fuzzy ...rough set theory is a powerful mathematical method for feature selection. The classical fuzzy rough set model is very sensitive to the noise while the noise samples in classification data often appear. In addition, fuzzy rough set theory does not fit well when the density distribution of the samples in the dataset varies greatly. Thus, it is of great significance to improve the robustness of fuzzy rough set models and its adaptability to data for feature selection. Inspired by these issues, we focus on the robust fuzzy rough set approach for feature selection. We first propose a robust fuzzy rough set model based on data distribution to achieve the purpose of anti-noise i.e., Noise-aware Fuzzy Rough Sets (NFRS) model. This model proposes a novel search mechanism, which weakens the sensitivity of the approximation operator to noise by considering the distribution of samples in the decision classes to weight the samples, further obtains three kinds of samples, i.e., intra-class samples, boundary samples, and outlier samples. Then, the degrees of relevance of the feature for class is defined by the dependency function based on the NFRS model to evaluate the significance of the feature subset. On this basis, an evaluation function about feature significance is constructed, which simultaneously considers the relevance and redundancy of a candidate feature provided for the selected subset and the remaining feature subset. A novel forward greedy search algorithm is presented to select a feature sequence. The selected features are subsequently evaluated with downstream classification tasks. Experimental using real-world datasets demonstrate the effectiveness of the proposed model and its superiority against comparison baseline methods.
Battery capacity estimation plays a crucial role in optimizing the performance and longevity of electric vehicles and stationary energy storage systems. However, accurately estimating battery ...capacity becomes challenging in real-world applications, particularly when dealing with unlabeled and noisy capacity data. To address this issue, this paper presents a co-learning framework incorporating both supervised and self-supervised learning for estimating battery capacity in the presence of few-labeled and noisy data. By training a shared encoder network with both self-supervised and supervised heads, the framework maximizes the agreement between the two heads in the latent space. The proposed approach demonstrates improved accuracy in battery capacity estimation in those challenging scenarios based on two public datasets. Comparative experiments show that the co-learning approach outperforms conventional end-to-end mapping methods, the average root mean square error of the proposed method is reduced at least by 36% and 19% under insufficient and noisy label conditions, respectively, leading to significant enhancements in estimation performance.
•A co-learning model is proposed for SOH estimation with unlabeled and noisy data.•The model integrates both supervised and self-supervised learning techniques.•The model accepts flexible inputs of partial charging data at various initial SOCs.•SOH estimation accuracy is significantly improved compared with the CNN model.
•We propose an iterative noise filter based on the fusion of classifiers.•A noisy score is introduced to control the noise sensitivity of the filter.•Our proposal removes less clean examples than ...other noise filters.•Our proposal enhances the performance of other noise filters considered.
In classification, noise may deteriorate the system performance and increase the complexity of the models built. In order to mitigate its consequences, several approaches have been proposed in the literature. Among them, noise filtering, which removes noisy examples from the training data, is one of the most used techniques. This paper proposes a new noise filtering method that combines several filtering strategies in order to increase the accuracy of the classification algorithms used after the filtering process. The filtering is based on the fusion of the predictions of several classifiers used to detect the presence of noise. We translate the idea behind multiple classifier systems, where the information gathered from different models is combined, to noise filtering. In this way, we consider the combination of classifiers instead of using only one to detect noise. Additionally, the proposed method follows an iterative noise filtering scheme that allows us to avoid the usage of detected noisy examples in each new iteration of the filtering process. Finally, we introduce a noisy score to control the filtering sensitivity, in such a way that the amount of noisy examples removed in each iteration can be adapted to the necessities of the practitioner. The first two strategies (use of multiple classifiers and iterative filtering) are used to improve the filtering accuracy, whereas the last one (the noisy score) controls the level of conservation of the filter removing potentially noisy examples. The validity of the proposed method is studied in an exhaustive experimental study. We compare the new filtering method against several state-of-the-art methods to deal with datasets with class noise and study their efficacy in three classifiers with different sensitivity to noise.
We describe in this article the techniques developed for the robust treatment of the static energy versus volume theoretical curve in the new version of the quasi-harmonic model code Comput. Phys. ...Commun. 158 (2004) 57. An average of strain polynomials is used to determine, as precisely as the input data allow it, the equilibrium properties and the derivatives of the static E(V) curve. The method provides a conservative estimation of the error bars associated to the fitting procedure. We have also developed the techniques required for detecting, and eventually removing, problematic data points and jumps in the E(V) curve. The fitting routines are offered as an independent octave package, called AsturFit, with an open source license.
Program title:AsturFit
Catalogue identifier: AEIY_v1_0
Program summary URL:http://cpc.cs.qub.ac.uk/summaries/AEIY_v1_0.html
Program obtainable from: CPC Program Library, Queenʼs University, Belfast, N. Ireland
Licensing provisions: GPL version 3
No. of lines in distributed program, including test data, etc.: 21 347
No. of bytes in distributed program, including test data, etc.: 620 496
Distribution format: tar.gz
Programming language: GNU Octave
Computer: Workstations
Operating system: Unix, GNU/Linux
Classification: 4.9
External routines: The GSL and OPTIM packages from the octaveforge site (http://octave.sourceforge.net/).
Nature of problem: Fit the total energy versus volume data of a solid to a continuous function and extract the equilibrium properties and the derivatives of the energy, with an estimation of the error introduced by the fitting procedure.
Solution method: The use of averages of strain polynomials allows a robust and reliable representation of the energy curve and its derivatives, together with a statistical estimation of the goodness of the calculated properties.
Additional comments: The techniques discussed have been implemented in Gibbs2, to be included with the second part of this article. Included here is the OCTAVE implementation of the routines, useful for interactive work and also for the creation of independent scripts. Some representative examples are included as test cases with a collection of data sets, test scripts, and model outputs.
Running time: Seconds at most in routine uses of the program. Special tasks like the bootstrap analysis may take up to some minutes.
► Robust fitting of energy versus volume curves using averages of strain polynomials. ► Error bars associated to the fits and thermodynamic properties. ► Detection of noise and problems in the input data. ► An octave package implementing the technique: Asturfit.
Since support vector regression (SVR) is a flexible regression algorithm, its computational complexity does not depend on the dimensionality of the input space, and it has excellent generalization ...capability. However, a central assumption with SVRs is that all the required data is available at the time of construction, which means these algorithms cannot be used with data streams. Incremental SVR has been offered as a potential solution, but its accuracy suffers with noise and learning speeds are slow. To overcome these two limitations, we propose a novel incremental regression algorithm, called online robust support vector regression (ORSVR). ORSVR solves nonparallel bound functions simultaneously. Hence, the large quadratic programming problem (QPP) in classical v-SVR are decomposed into two smaller QPPs. An incremental learning algorithm then solves each QPP step-by-step. The results of a series of comparative experiments demonstrate that the ORSVR algorithm efficiently solves regression problems in data streams, with or without noise, and speeds up the learning process.
Identifiability in robust estimation of tree structured models Casanellas, Marta; Garrote-López, Marina; Zwiernik, Piotr
Bernoulli : official journal of the Bernoulli Society for Mathematical Statistics and Probability,
2/2024, Letnik:
30, Številka:
1
Journal Article