An Evaluation of Outlier Detection Using Machine Learning in Medicine

Abstract

Detecting outliers is a serious issue that has been investigated in a number of academic and application fields. To effectively detect outliers, researchers are still developing reliable systems. In the area of medical research, the outlier detection problem has several relevant applications. Big data is rapidly being acknowledged and valued by individuals as a result of the extraordinarily quick increase of data in numerous industries. Numerous parties have taken an interest in medical big data, which can best illustrate the utility of big data. Numerous notes about a patient's ailments, therapies, and lab results can be found in their medical records. typically incorporate several different sorts of data and generate a lot of information. These databases can offer crucial data to enhance hospital management and clinical decision-making. Some specifics discovered in medical databases are rarely present in other non-medical sources. In this situation, anomalous patterns in health records (such as issues with data quality) can be found using outlier detection techniques, which will lead to better data and knowledge for decision-making.

In the field of data analysis, outlier detection has long been a key idea. The direct relationship between data outliers and real-world abnormalities, which are of significant interest to analysts, has recently been realised in a number of application domains. Finding patterns in data that deviate from predicted typical behaviour is known as outlier detection. The secondary sources are where the study's data was gathered. The primary goal of this study is to examine outlier detection in medical data using machine learning techniques.