Dealing with overlap and imbalance: a new metric and approach

Borsos, Zalán; Lemnaru, Camelia; Potolea, Rodica

doi:10.1007/s10044-016-0583-6

Cited by 19 publications

(20 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this section, we compared ROS and ROA algorithms results with those of RONS and Standard method (STD). As already explained, all parameters are maintained same with the objective of formula (11) where Raug value is used to measure the decline of overlapping degree. We have considered the standard method which represented the standard classification without using feature selection or resampling technique.…”

Section: Resultsmentioning

confidence: 99%

“…Based on this method, the augmented R-value has been introduced to measure the overlap degree in order to minimize it (see formula (2 which used both equations ( 1) and (3))). The following is given to define R-aug [11]:…”

Section: Proposed Algorithmsmentioning

confidence: 99%

“…In fact,R aug is mainly based on R(C i ), which indicates the R-value of class C i already proposed in reference [11].…”

Section: Proposed Algorithmsmentioning

confidence: 99%

“…This measurement can be used to assess the overlap degree of class-imbalanced data. Larger the value is, higher overlap degree of a data set is [11]. In this study, we focus on improving the classification performance in the settings of class-imbalance learning.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Minimizing the Overlapping Degree to Improve Class-Imbalanced Learning Under Sparse Feature Selection: Application to Fraud Detection

et al. 2021

View full text Add to dashboard Cite

In recent years, the classification of class-imbalanced data has obtained increasing attention across different scientific areas such as fraud detection, metabolomics, Cancer diagnosis, etc. This interest comes after proving the negative effect of overlapping on the performance of class-imbalanced learning. Based on augmented R-value, our proposed strategy aims to select features that make data achieve the minimal overlap degree, so improving the performance of classification as well. In this context, we present three feature selection algorithms RONS (Reduce Overlapping with No-sampling), ROS (Reduce Overlapping with SMOTE), and ROA (Reduce Overlapping with ADASYN), which are built through sparse feature selection to minimize the overlapping and perform binary classification. Also, a re-sampling process has been included in both ROS and ROA. Simulation results show that our proposed algorithms as feature selection methods manage the variation of false discovery rate during the selection of main features for the process modeling. For the experiment, four credit card datasets have been selected to test the performance of our algorithms. Using F-measure and Gmean evaluation metrics, the results reveal that our proposed algorithms are considerably recommended compared with classical feature selection methods. Besides, this effective feature selection strategy can be extended as an alternative to deal with class-imbalanced learning problems that involve overlapping.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Proposed Algorithmsmentioning

confidence: 99%

“…In fact,R aug is mainly based on R(C i ), which indicates the R-value of class C i already proposed in reference [11].…”

Section: Proposed Algorithmsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Minimizing the Overlapping Degree to Improve Class-Imbalanced Learning Under Sparse Feature Selection: Application to Fraud Detection

et al. 2021

View full text Add to dashboard Cite

show abstract

“…Machine learning [ 23 ], according to the specialized literature, includes both unsupervised learning [ 24 ] and supervised learning. Supervised learning [ 25 ] is based on a training model that uses a set of labeled data. This model is then used to map between a series of new observations and their category based on a function that generates it at the time of training.…”

Section: Software Application Developmentmentioning

confidence: 99%

Development of a Virtual Reality Simulator for an Intelligent Robotic System Used in Ankle Rehabilitation

Covaciu

Pîslă

Iordan

2021

Sensors

View full text Add to dashboard Cite

The traditional systems used in the physiotherapy rehabilitation process are evolving towards more advanced systems that use virtual reality (VR) environments so that the patient in the rehabilitation process can perform various exercises in an interactive way, thus improving the patient’s motivation and reducing the therapist’s work. The paper presents a VR simulator for an intelligent robotic system of physiotherapeutic rehabilitation of the ankle of a person who has had a stroke. This simulator can interact with a real human subject by attaching a sensor that contains a gyroscope and accelerometer to identify the position and acceleration of foot movement on three axes. An electromyography (EMG) sensor is also attached to the patient’s leg muscles to measure muscle activity because a patient who is in a worse condition has weaker muscle activity. The data collected from the sensors are taken by an intelligent module that uses machine learning to create new levels of exercise and control of the robotic rehabilitation structure of the virtual environment. Starting from these objectives, the virtual reality simulator created will have a low dependence on the therapist, this being the main improvement over other simulators already created for this purpose.

show abstract

A Hybrid Model Based on Samples Difficulty for Imbalanced Data Classification

Shan,

Chung

2023

Artificial Neural Networks and Machine Learning – ICANN 2023

View full text Add to dashboard Cite

Dealing with overlap and imbalance: a new metric and approach

Cited by 19 publications

References 41 publications

Minimizing the Overlapping Degree to Improve Class-Imbalanced Learning Under Sparse Feature Selection: Application to Fraud Detection

Minimizing the Overlapping Degree to Improve Class-Imbalanced Learning Under Sparse Feature Selection: Application to Fraud Detection

Development of a Virtual Reality Simulator for an Intelligent Robotic System Used in Ankle Rehabilitation

A Hybrid Model Based on Samples Difficulty for Imbalanced Data Classification

Contact Info

Product

Resources

About