Open AccessFeature Selection Methods for Predicting Household Food Insecurity
Author(s) -
Dorsewamy,
Mersha Nigus
Publication year - 2020
Publication title -
international journal of recent technology and engineering
Language(s) - English
Resource type - Journals
ISSN - 2277-3878
DOI - 10.35940/ijrte.a2382.059120
Subject(s) - feature selection , support vector machine , naive bayes classifier , random forest , feature (linguistics) , dimensionality reduction , computer science , artificial intelligence , univariate , pattern recognition (psychology) , data mining , selection (genetic algorithm) , k nearest neighbors algorithm , logistic regression , filter (signal processing) , machine learning , multivariate statistics , philosophy , linguistics , computer vision
Feature selection is a method of dimension reduction that is used to select a specific subset of appropriate features from the original features by removing unnecessary and redundant features that do not have a benefit in classification or prediction. In this paper, the feature selection approach was conducted using three feature selection methods namely: Filter based, Wrapper based and Embedded based to predict household food insecurity from the household income, consumption, and expenditure survey data (HICE). To implement the above feature selection methods, we proposed new hybrid method by integrating the filter based feature selection methods which is Feature importance, Univariate (chi-square) and Correlation coefficient. To validate the efficiency of the proposed feature selection methods, we used five classification algorithms namely: K-Nearest Neighbor (KNN), Logistic Regression (LR), Support Vector Machine (SVM), Random Forest (RF), and Naive Bayes (NB).