An outliers detection and elimination framework in classification task of data mining
Article Type
Research Article
Publication Title
Decision Analytics Journal
Abstract
An outlier is a datum that is far from other data points in which it occurs. It can have a considerable impact on the output. Therefore, removing or resolving it before the analysis is essential to prevent skewing. Outliers in a survey sampling can have a significant outcome on statistical results. The goal of discovering outliers in data mining is to find a pattern in data that does not conform to expected behavior. In this paper, we have proposed a framework in which a popular statistical approach termed Inter-Quartile Range (IQR) is used to detect outliers in data and deal with them by Winsorizing method. A radial basis function network trained by teaching a learning-based optimization model is developed using the preprocessed dataset under this framework. A few standard University of California Irvine (UCI)datasets are employed to measure the framework's effectiveness. The outcome of the experiments shows that our proposed framework can be a viable alternative tool for the classification task of data mining where prior outliers preprocessing is necessary.
DOI
https://10.1016/j.dajour.2023.100164
Publication Date
3-1-2023
Recommended Citation
Dash, Ch Sanjeev Kumar; Behera, Ajit Kumar; Dehuri, Satchidananda; and Ghosh, Ashish, "An outliers detection and elimination framework in classification task of data mining" (2023). Journal Articles. 3824.
https://digitalcommons.isical.ac.in/journal-articles/3824
Comments
Open Access, Gold