Deep Ensemble Outlier Detection in Lymphography Dataset

Abstract

Important data instances known as outliers are those whose characteristics differ from those of the majority of the instances in a dataset. With a variety of uses, including fraud detection in credit card transactions and intrusion detection in computer communications, outlier detection is a crucial field of study in statistics and data mining. In the medical field, diseases can also be identified from a variety of laboratory reports using outlier detection techniques. Researchers have developed a number of techniques to identify outliers in healthcare systems. This research aims to identify the outliers using Deep Ensemble Approach in lymphography dataset. In order to identify outliers in the lymphography dataset, this research suggests an ensemble approach based on deep learning and isolation forests. The suggested approach outperforms the individual models, according to experimental results using the lymphography dataset from the UCI machine learning repository that is accessible to the general public.