AN EFFICIENT FEATURE SELECTION AND CLASSIFICATION FOR THE CROP FIELD IDENTIFICATION: A HYBRIDIZED WRAPPER BASED APPROACH

Abstract

Agricultural stakeholders are concerned about the anticipated crop production
before the harvest. Many countries throughout the world employ computational technique's
for predicting yield ahead of harvest to assess a country's food security and issue warnings
about impending food shortages. This is a common method that aids strategy planners and
decision-makers, particularly in rural economies. Crop statistical models have been used to
track crop development and forecast production. The only inputs available at the field level
will yield a prediction for a narrow region; remote sensing observations cover a broad area.
They may be repeated at regular intervals, allowing for large-scale crop modelling. Crop
yield prediction study necessitates a variety of production parameters and algorithms. Some
algorithms are used to determine the optimum feature subset for improved prediction, while
others are used to determine prediction. The proposed Correlation based Sequential Forward
Feature Selection (CSFFS) is compared with the existing feature selection approaches. The
classification with proposed feature selection attains effective accuracy in crop prediction.