A MULTICRITERIA INDEX USING NEURAL NETWORK TO EVALUATE THE POTENTIAL LANDS OF MAIZE

Abstract

The criteria for planting maize should be consistent with sensible and ecological criteria to determine the potential lands. However, there is still a lack of proven methodology for this evaluation. The purpose of this analysis was to determine the parameters that affect the multi-criteria decision of maize, with the aim of a new method on the land suitability analysis. The land suitability analysis proposed was based on GIS-analysis and management parameters such as distance from roads, rivers, slope, LULC, elevation, soil type, NDVI, SAVI, rainfall, and temperature. We have found a sample of 4590 maize in Tuban, East Java, Indonesia. Based on the above criteria, maize has been classified into four groups according to FAO. Moreover, we analyzed was done using Neural Network. Results showed that the integrated AHP with Neural Network to evaluate the lands inferred that 66.7 percent of the study area was classified as highly suitable, 30.2 percent were moderately suitable, and 3 percent were marginally suitable for Maize Cultivation in Tuban Regency. The approach presented in this analysis can be extended in this analysis can be extended to other maize areas also other crops as a decision-making system.