Recursive Feature Elimination with Feature Ranking-Based Feature Selection for Healthcare Data
Author(s):
Arati Kale, Dr.Dev Ras Pandey
Keywords:
Feature Selection, Recursive Feature Elimination, Ranking, health data, machine learning
Abstract
The feature selection method is a learning acquisition technique that selects the most pertinent features. Feature selection aims to reduce computational overload and improve the classifier by enumerating essential features. The presence of replicate and unimportant features in an extensive data set can dramatically reduce the efficiency of machine learning models. This paper proposes an efficient feature selection algorithm for healthcare data using recursive feature elimination with feature ranking (RFE_FR) to improve classification performance. The different feature ranking methods, including info gain, relief, and correlation-based ranking approaches, are used to compute the feature rank. It reduces the feature dimension and selects features efficiently. This approach improves performance by removing associated and redundant features from the dataset. To investigate the RFE_FR performance, the machine learning classification algorithms KNN, NB, RF, and SVM are employed. The accuracy of the suggested method performance is assessed using the real-time healthcare dataset. This study demonstrates that, when comparing the model efficiency with and without feature selection, the RFE_FR strategies selected have a significant favourable effect on the model performance.
Article Details
Unique Paper ID: 162237
Publication Volume & Issue: Volume 10, Issue 8
Page(s): 474 - 480
Article Preview & Download
Share This Article
Join our RMS
Conference Alert
NCSEM 2024
National Conference on Sustainable Engineering and Management - 2024