DOI: 10.14704/NQ.2022.20.11.NQ66043

Agro Informatic Data Classification using Machine Learning Techniques

K. Uma maheswari,G. Shobana

Abstract

Agro-informatics is one of the newly emerging fields where machine learning and data analytics are applied to the agricultural data to assess grain quality, predict crop disease at an earlier stage, classify intra-species of crops and help in the analysis of several other agriculture related issues. Manual investigation of crops for its quality, size and disease is a tedious time-consuming process. The conventional methods also involve high production costs. Advanced technology involves automatic image capturing, data acquisition, pre-processing and application of machine learning for the analysis. In this paper, dry-beans data are classified using a proposed feature reduction framework using Principal Component Analysis and Linear Discriminant Analysis with machine learning algorithms. Multi-layer perceptron, Random Forest and Support Vector Machine performed well with 93% for multinomial classification while 98% for binary classification. With balanced and comprehensive groups of two bean samples, binary classification achieved higher accuracy than multinomial classification with seven groups of beans.

Keywords

Agro-informatic; dry-bean; multi-layer perceptron; random forest;support vector machine.

Full Text

References