News

Real data on the basis of researchers try to build decision models and draw conclusions is often not perfect, which means that they require costly and time-consuming preprocessing as well as the selection of appropriate methods of discovering the knowledge contained in them.

One of the main problems facing researchers in this area is the uneven distribution of data between classes. It can be observed in many areas, such as medical diagnosis of rare diseases or fraud detection in banking transactions. In most of the mentioned cases, the identification of rare objects is crucial. Unfortunately, the uneven distribution of classes in the dataset causes problems in constructing machine learning models that would be able to classify such objects correctly.

Another common problem is the curse of dimensionality. In many cases, many parameters describing the analyzed objects increase the computational complexity and make it difficult to generalize the decision model. Therefore, it is often necessary to use dimensionality reduction techniques.

Session IS29 "(DECREASE) Delving into Imbalanced, Overlapping and Multidimensional Data: Recent Advances and Challenges," held as part of the International Conference on Knowledge-Based and Intelligent Information & Engineering Systems KES2023 (Core-B, 70 pts. ) aims to enable researchers to exchange knowledge and experience in solving a wide range of problems related to the imperfection of the analyzed data, including their imbalance, multidimensionality, noise, the existence of outliers or, for example, class overlap.

More information about the session and the entire Conference can be found at: http://kes2023.kesinternational.org/cms/userfiles/is29.pdf and http://kes2023.kesinternational.org/.

We cordially invite you to submit papers for the IS29 session. Submissions are accepted until April 3, 2023.