Document is current

Article History

Received: 17 July 2023

Accepted: 17 October 2024

First Online: 7 January 2025

Declarations

:

: Our method detects outliers regardless of data model and data label. Our method considers both local and global features at the same time in order to detect outliers. We use adaptive extended neighbors to calculate distance and density measures. We provide and combine three different and effective expressions to improve the performance of the proposed algorithm. Each of them focuses on certain problem and solves it. In this method, we consider closer neighbors more than further neighbors. So, by using this concept we eliminate sensitivity to values. Our results show higher accuracy in real and artificial datasets compared to the proposed methods in this field. Also, it has the least sensitivity to parameters, and on the other hand, it has not increased the time complexity.