IoT based Smart Farming : Feature subset selection for optimized high-dimensional data using improved GA based approach for ELM

•Uncertain data classification with optimal features for high dimensional dataset.•Improved Genetic Algorithm based multilevel parameter optimization for ELM (IGA-ELM).•Smart Farming DSS – clear separation of plant disease and nutrient deficiency.•Maximizing classification accuracy by minimizing the...

Full description

Saved in:
Bibliographic Details
Published in:Computers and electronics in agriculture 2019-06, Vol.161, p.225-232
Main Authors: Kale, Archana P., Sonavane, Shefali P.
Format: Article
Language:English
Subjects:
Agriculture
Agronomy
Classification
Computer simulation
Datasets
Extreme learning machine
Farming
Feature subset selection problem
Genetic algorithms
Internet of Things
Optimization
Organic chemistry
Pattern classification problem
Plant disease detection
Plant diseases
Support systems
Uncertainty data
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:•Uncertain data classification with optimal features for high dimensional dataset.•Improved Genetic Algorithm based multilevel parameter optimization for ELM (IGA-ELM).•Smart Farming DSS – clear separation of plant disease and nutrient deficiency.•Maximizing classification accuracy by minimizing the number of features. Agriculture is one of the major backbones of Indian economy where around 60% of people are depending directly or indirectly upon agriculture. The expert advice is required for distinguishing the plant disease damage and nutrient imbalance. It is observed that, the conventional judgmental analysis is not enough while deciding the quantity of chemical or fertilizer to be used. The mis-proportional dose harms the health of the crop and hence the living beings. To overcome the said problem, this paper proposes an Internet of things (IoT) based Smart Farming decision support system with an improved genetic algorithm (IGA) based multilevel parameter optimized feature selection algorithm for ELM classifier (IGA-ELM). The proposed work is applied to benchmark high dimensional biomedical datasets as well as for real time applications (plant disease dataset) which provides 9.52% and 5.71% improvement in the classification accuracy by reducing 58.50% and 72.73% features respectively. Simulation results demonstrate that IGA-ELM has the capability to handle optimization, uncertainty and supervised binary classification problems with improved classification accuracy even though reduced the number of features.
ISSN:0168-1699
1872-7107
DOI:10.1016/j.compag.2018.04.027