各位同仁们好，我们近期在《
Applied Soft Computing
》（
JCR
一区
,
中科院二区，影响因子
8.263
） 期刊上组织了关于
“
Computational Intelligence for Unbalanced Classification
”
的专刊。该专刊将收录有关将
计算智能相关算法（
包括演化计算、模态逻辑、以及神经网络
），
应用于解决不平衡数据分类的论文。
相关应用
包括医疗诊断、故障检测、异常检测等。
【投稿地址】
：
https://ees.elsevier.com/asoc
【
投稿时请选择
】：
VSI: Unbalanced Classif
作为投稿类型（
系统
4
月
1
日开放
）
欢迎各位专家学者、学生们踊跃投稿！
谢谢！
Special issue information:
In many real-world applications, such as fault detection, it is usually hard to avoid unbalanced data, which exhibits a skewed distribution between its classes. However, many standard classification algorithms from machine learning do not specifically consider the skewed data distribution, and treat all the instances in a dataset as being equally important. If the class imbalance issue is not well-addressed, classifiers are likely to ignore the class of interest which is constituted by a few instances.
In order to address the class imbalance issue, many techniques/methods in Machine Learning have been developed to either rebalance data or improve standard classification algorithms. Sampling methods are most well-known methods at the data level, with the aim to rebalance data for guaranteeing the same or a similar number of instances per class. There are a class of dynamic sampling techniques that could be regarded as some kind of implicit sampling or weighting of training instances. Cost-sensitive learning is popular in the algorithm-level category of methods, which takes costs (mostly misclassification costs) into account. However, these existing methods may become less effective when the amount of data is big or/and various types of data are intertwined with each other.
Computational intelligence
is an important sub-field of artificial intelligence, which covers three research branches, i.e., Evolutionary Computation, Fuzzy Sets, and Artificial Neural Networks. Techniques from all the three branches have been applied and achieved great contributions to a wide range of challenging tasks in unbalanced classification. In the big data era, there are many new challenging real-world tasks involved with more complicated unbalanced data, such as high-dimensional unbalanced data and streaming unbalanced data. This arises new challenges to existing methods for maintaining their effectiveness under the conditions of limited domain knowledge, human efforts, and computational resources available. Therefore, new approaches that address those challenges are desired.
The aim of this special issue is to promote the contemporary use of theories and methodologies from different branches in computational intelligence for addressing challenging tasks and open questions in unbalanced classification.
TOPICS
：
We welcome studies and contributions that propose novel methods based on different computational intelligence paradigms for unbalanced classification and its applications. Topics of interest include but are not limited to:
-
Undersampling for Unbalanced Classification
-
Oversampling for Unbalanced Classification
-
Hybrid Sampling for Unbalanced Classification
-
Cost-sensitive Learning for Unbalanced Classification
-
Active Learning for Unbalanced Streaming Data Classification
-
Ensemble Learning for Unbalanced Data Classification
-
Fuzzy Rule-based Classification Systems for Unbalanced Classification
-
Rough sets for Unbalanced Classification
-
Deep Neural Networks for Unbalanced Classification
-
Interpretability of Over-complicated Models in Unbalanced Classification
-
Multi-task Learning for Unbalanced Classification
-
Transfer Learning for Unbalanced Classification
-
Feature Selection/Construction/Extraction/Ranking/Analysis for Unbalanced Classification with High-dimensional Data
-
Real-world Applications of Unbalanced Classification, e.g., Medical Diagnosis, Fault Detection, Anomaly Detection, and Text Mining
Submission Instructions
：
Paper submissions for the special issue should follow the submission format and guidelines for regular papers and be submitted at https://ees.elsevier.com/asoc. Each submission must contribute to “Computational Intelligence for Unbalanced Classification”. Papers that either lack originality, clarity in presentation, or fall outside the scope of the special issue will be desk-rejected without further review. Authors should select “VSI: Unbalanced Classif.” when they reach the “Article Type” step in the submission process. The submitted papers must propose original research that has not been published nor is currently under review in other venues.
