An FCR Approach Towards Detection of Outliers for Medical Data

An FCR Approach Towards Detection of Outliers for Medical Data

Regression analysis has been brought into practical functionality in this paper to intensify the magnitude of adequacy for outliers' detection in medical data. Previously, linear regression model for the outlier detection in medical data has been used. The use of linear model to detect outliers...

Full description

Saved in:
Bibliographic Details
Published in:2020 IEEE 17th International Conference on Smart Communities: Improving Quality of Life Using ICT, IoT and AI (HONET) pp. 224 - 230
Main Authors: Iqbal, Sidra, Ajmeri, Hafiz Bahloul, Bibi, Sumaira, Wahid, Abdul
Format: Conference Proceeding
Language:English
Published: IEEE 14-12-2020
Subjects:
Anomaly detection
Data models
detecting anomalies
detection
Electrical engineering
FCR approach
fitted cubic regression
linear regression
Mathematical model
medical administrative data processing
medical data
Outlier
outlier detection
Predictive models
regression analysis
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Regression analysis has been brought into practical functionality in this paper to intensify the magnitude of adequacy for outliers' detection in medical data. Previously, linear regression model for the outlier detection in medical data has been used. The use of linear model to detect outliers leaves a certain amount of gap in increasing the efficiency of the expected values. Linear regression basically represents linear relationship between data and sometimes loosely fits the data. The proposed FCR technique on contrary contributes towards increasing the efficiency of expected data and improved effectiveness towards detecting outliers. The proposed method is adept to best fit curve to the data. It further verifies to be a constructive asset that increases the model adequacy on medical data. The proposed approach in this paper has demonstrated to deliver enhanced adequacy paralleled to linear regression.
ISSN:1949-4106
DOI:10.1109/HONET50430.2020.9322843