scholarly journals Mitigating Data Imbalance Problem in Transformer-Based Intent Detection

2022 ◽  
Author(s):  
Osman BÜYÜK ◽  
Mustafa ERDEN ◽  
Levent ARSLAN
Keyword(s):  
Data Imbalance ◽  
Imbalance Problem

Improving Classification Accuracy on Imbalanced Data by Ensembling Technique

2017 ◽  
Vol 19 (1) ◽  
pp. 42-49
Author(s):  
Divya Agrawal ◽  
Padma Bonde
Keyword(s):  
Classification Accuracy ◽  
Area Under The Curve ◽  
Imbalanced Data ◽  
Data Handling ◽  
Imbalanced Dataset ◽  
Classification Techniques ◽  
Data Imbalance ◽  
Logistics Regression ◽  
Good Evaluation ◽  
Imbalance Problem

Prediction using classification techniques is one of the fundamental feature widely applied in various fields. Classification accuracy is still a great challenge due to data imbalance problem. The increased volume of data is also posing a challenge for data handling and prediction, particularly when technology is used as the interface between customers and the company. As the data imbalance increases it directly affects the classification accuracy of the entire system. AUC (area under the curve) and lift proved to be good evaluation metrics. Classification techniques help to improve classification accuracy, but in case of imbalanced dataset classification accuracy does not predict well and other techniques, such as oversampling needs to be resorted. Paper presented Voting based ensembling technique to improve classification accuracy in case of imbalanced data. The voting based ensemble is based on taking the votes on the best class obtained by the three classification techniques, namely, Logistics Regression, Classification Trees and Discriminant Analysis. The observed result revealed improvement in classification accuracy by using voting ensembling technique.

scholarly journals Analysis of Factors Affecting Hit-and-Run and Non-Hit-and-Run in Vehicle-Bicycle Crashes: A Non-Parametric Approach Incorporating Data Imbalance Treatment

2019 ◽  
Vol 11 (5) ◽  
pp. 1327 ◽  
Author(s):  
Bei Zhou ◽  
Zongzhi Li ◽  
Shengrui Zhang ◽  
Xinfen Zhang ◽  
Xin Liu ◽  
...  
Keyword(s):  
Geometric Mean ◽  
Regression Tree ◽  
Classification Performance ◽  
Classification And Regression Tree ◽  
Factors Affecting ◽  
Data Imbalance ◽  
Imbalance Problem ◽  
Hit And Run ◽  
Reported Data ◽  
Cart Model

Hit-and-run (HR) crashes refer to crashes involving drivers of the offending vehicle fleeing incident scenes without aiding the possible victims or informing authorities for emergency medical services. This paper aims at identifying significant predictors of HR and non-hit-and-run (NHR) in vehicle-bicycle crashes based on the classification and regression tree (CART) method. An oversampling technique is applied to deal with the data imbalance problem, where the number of minority instances (HR crash) is much lower than that of the majority instances (NHR crash). The police-reported data within City of Chicago from September 2017 to August 2018 is collected. The G-mean (geometric mean) is used to evaluate the classification performance. Results indicate that, compared with original CART model, the G-mean of CART model incorporating data imbalance treatment is increased from 23% to 61% by 171%. The decision tree reveals that the following five variables play the most important roles in classifying HR and NHR in vehicle-bicycle crashes: Driver age, bicyclist safety equipment, driver action, trafficway type, and gender of drivers. Several countermeasures are recommended accordingly. The current study demonstrates that, by incorporating data imbalance treatment, the CART method could provide much more robust classification results.

scholarly journals Improvement of P300-Based Brain–Computer Interfaces for Home Appliances Control by Data Balancing Techniques

Sensors ◽  
2020 ◽  
Vol 20 (19) ◽  
pp. 5576
Author(s):  
Taejun Lee ◽  
Minju Kim ◽  
Sung-Phil Kim
Keyword(s):  
Information Transfer ◽  
Regularization Parameter ◽  
Support Vector ◽  
Oddball Paradigm ◽  
Svm Classifier ◽  
Sampling Techniques ◽  
Brain Computer Interfaces ◽  
Data Imbalance ◽  
Imbalance Problem ◽  
Computer Interfaces

The oddball paradigm used in P300-based brain–computer interfaces (BCIs) intrinsically poses the issue of data imbalance between target stimuli and nontarget stimuli. Data imbalance can cause overfitting problems and, consequently, poor classification performance. The purpose of this study is to improve BCI performance by solving this data imbalance problem with sampling techniques. The sampling techniques were applied to BCI data in 15 subjects controlling a door lock, 15 subjects an electric light, and 14 subjects a Bluetooth speaker. We explored two categories of sampling techniques: oversampling and undersampling. Oversampling techniques, including random oversampling, synthetic minority oversampling technique (SMOTE), borderline-SMOTE, support vector machine (SVM) SMOTE, and adaptive synthetic sampling, were used to increase the number of samples for the class of target stimuli. Undersampling techniques, including random undersampling, neighborhood cleaning rule, Tomek’s links, and weighted undersampling bagging, were used to reduce the class size of nontarget stimuli. The over- or undersampled data were classified by an SVM classifier. Overall, some oversampling techniques improved BCI performance while undersampling techniques often degraded performance. Particularly, using borderline-SMOTE yielded the highest accuracy (87.27%) and information transfer rate (8.82 bpm) across all three appliances. Moreover, borderline-SMOTE led to performance improvement, especially for poor performers. A further analysis showed that borderline-SMOTE improved SVM by generating more support vectors within the target class and enlarging margins. However, there was no difference in the accuracy between borderline-SMOTE and the method of applying the weighted regularization parameter of the SVM. Our results suggest that although oversampling improves performance of P300-based BCIs, it is not just the effect of the oversampling techniques, but rather the effect of solving the data imbalance problem.

Re-Weighted Interval Loss for Handling Data Imbalance Problem of End-to-End Keyword Spotting

2020 ◽  
Author(s):  
Kun Zhang ◽  
Zhiyong Wu ◽  
Daode Yuan ◽  
Jian Luan ◽  
Jia Jia ◽  
...  
Keyword(s):  
Keyword Spotting ◽  
Data Imbalance ◽  
Imbalance Problem ◽  
End To End

Two-Stage Neural Network Classifier for the Data Imbalance Problem with Application to Hotspot Detection

2021 ◽  
Author(s):  
Bingshu Wang ◽  
Lanfan Jiang ◽  
Wenxing Zhu ◽  
Longkun Guo ◽  
Jianli Chen ◽  
...  
Keyword(s):  
Neural Network ◽  
Neural Network Classifier ◽  
Hotspot Detection ◽  
Two Stage ◽  
Data Imbalance ◽  
Imbalance Problem

scholarly journals A Novel SMOTE-Based Classification Approach to Online Data Imbalance Problem

2016 ◽  
Vol 2016 ◽  
pp. 1-14 ◽  
Author(s):  
Chunlin Gong ◽  
Liangxian Gu
Keyword(s):  
Learning Algorithm ◽  
Sampling Strategy ◽  
Training Sample ◽  
Classification Performance ◽  
Online Data ◽  
Classification Approach ◽  
Data Imbalance ◽  
Imbalance Problem ◽  
Efficient Learning ◽  
Sample Set

In many practical engineering applications, data are usually collected in online pattern. However, if the classes of these data are severely imbalanced, the classification performance will be restricted. In this paper, a novel classification approach is proposed to solve the online data imbalance problem by integrating a fast and efficient learning algorithm, that is, Extreme Learning Machine (ELM), and a typical sampling strategy, that is, the synthetic minority oversampling technique (SMOTE). To reduce the severe imbalance, the granulation division for major-class samples is made according to the samples’ distribution characteristic, and the original samples are replaced by the obtained granule core to prepare a balanced sample set. In online stage, we firstly make granulation division for minor-class and then conduct oversampling using SMOTE in the region around granule core and granule border. Therefore, the training sample set is gradually balanced and the online ELM model is dynamically updated. We also theoretically introduce fuzzy information entropy to prove that the proposed approach has the lower bound of model reliability after undersampling. Numerical experiments are conducted on two different kinds of datasets, and the results demonstrate that the proposed approach outperforms some state-of-the-art methods in terms of the generalization performance and numerical stability.

Sign in / Sign up

Export Citation Format

Share Document