Cost Sensitive Time-Series Classification

Roychoudhury, Shoumik; Ghalwash, Mohamed; Obradović, Zoran

doi:10.1007/978-3-319-71246-8_30

Cited by 12 publications

(15 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As in LTSfAUC, existing shapelet methods such as Refs. 1,3 that do not optimize the pAUC cannot discover shapelets that discriminate minor negative instances. As a result, their pAUC also degrades.…”

Section: Ltspauc 393mentioning

confidence: 99%

“…2 In recent years, several learning-based shapelet methods have been proposed. 1,3 These learning-based methods use stochastic gradient descent (SGD) algorithms to learn both shapelets and classifiers. Thus, learned shapelets are not restricted to being subseries in training timeseries instances.…”

Section: Introductionmentioning

confidence: 99%

“…4,5 Furthermore, they improve classification performance in terms of accuracy and F-measures. 1,3 As shown in Figure 1, graphing the true-positive rate (TPR) against the false-positive rate (FPR) as the decision threshold is varied yields the receiver operating characteristic (ROC) curve. Here, TPR is the probability of correctly classifying a positive instance, and FPR is the probability of incorrectly classifying a negative instance as positive.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

LTSpAUC: Learning Time-Series Shapelets for Partial AUC Maximization

et al. 2020

View full text Add to dashboard Cite

Shapelets are discriminative segments used to classify time-series instances. Shapelet methods that jointly learn both classifiers and shapelets have been studied in recent years because such methods provide both interpretable results and superior accuracy. The partial area under the receiver operating characteristic curve (pAUC) for a low range of false-positive rates (FPR) is an important performance measure for practical cases in industries such as medicine, manufacturing, and maintenance. In this article, we propose a method that jointly learns both shapelets and a classifier for pAUC optimization in any FPR range, including the full AUC. In addition, we propose the following two extensions for shapelet methods: (1) reducing algorithmic complexity in time-series length to linear time and (2) explicitly determining the classes that shapelets tend to match. Comparing with state-of-theart learning-based shapelet methods, we demonstrated the superiority of pAUC on UCR time-series data sets and its effectiveness in industrial case studies from medicine, manufacturing, and maintenance.

show abstract

Section: Ltspauc 393mentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

LTSpAUC: Learning Time-Series Shapelets for Partial AUC Maximization

et al. 2020

View full text Add to dashboard Cite

show abstract

“…In data manipulation level [2,10,[13][14][15], time-series datasets were re-established through over-sampling of positive samples, or under-sampling of negative samples, or both. In algorithmic modification level [16], classifiers were modified by predefining higher costs or class weights for false positive samples. However, there were some problems of those two levels approaches needed to be noticed.…”

Section: Introductionmentioning

confidence: 99%

“…Algorithmic modification approaches need predefining the cost weight or cost matrix and the exact settings are difficultly found. Besides, most of two levels methods applied algorithms like KNN-DTM [17], SVM [13,15], Shaplets [16]. Those classic algorithms need heavy hand-crafted works on data preprocessing or feature engineering and they are not appropriate for large volume dataset.…”

Section: Introductionmentioning

confidence: 99%

Cost-sensitive convolutional neural networks for imbalanced time series classification

Geng¹,

Luo²

2019

IDA

View full text Add to dashboard Cite

Some deep convolutional neural networks were proposed for time-series classification and class imbalanced problems. However, those models performed degraded and even failed to recognize the minority class of an imbalanced temporal sequences dataset. Minority samples would bring troubles for temporal deep learning classifiers due to the equal treatments of majority and minority class. Until recently, there were few works applying deep learning on imbalanced time-series classification (ITSC) tasks. Here, this paper aimed at tackling ITSC problems with deep learning. An adaptive cost-sensitive learning strategy was proposed to modify temporal deep learning models. Through the proposed strategy, classifiers could automatically assign misclassification penalties to each class. In the experimental section, the proposed method was utilized to modify five neural networks. They were evaluated on a large volume, real-life and imbalanced time-series dataset with six metrics. Each single network was also tested alone and combined with several mainstream data samplers. Experimental results illustrated that the proposed costsensitive modified networks worked well on ITSC tasks. Compared to other methods, the cost-sensitive convolution neural network and residual network won out in the terms of all metrics. Consequently, the proposed cost-sensitive learning strategy can be used to modify deep learning classifiers from cost-insensitive to costsensitive. Those cost-sensitive convolutional networks can be effectively applied to address ITSC issues.

show abstract