Crop Recommendation System for Precision Agriculture Using Fuzzy Clustering Based Ant Colony Optimization

Abstract

An application that also involves machine learning to suggest the crop variety and also to predict the yield. To be precise in predicting the crops, the project analyzes the nutrients that are naturally present in the soil and the crop productivity based on the location. It compares the accuracy obtained by various network learning techniques and the most appropriate result will be given to the end-user. So, the project aims to predict the amount of crop yield in the desired area in the ratio of tons per hectare. This paper proposes the fuzzy clustering designed by the Fuzzy Clustering Ant Colony Optimization (ACO) algorithm (FCACO). FCACO’s main objective is to develop the performance and effectiveness of the fuzzy cluster design. In the FCACO, the fuzzy clustering structure includes the count of fuzzy sets and rules in each of the input variables that are created online by the fuzzy clustering. Distinguished by the type of classical score grid, a previous part is partitioned flexibly, and the phenomenon of strongly nested fuzzy sets is avoided. After, a new rule is generated, and from the candidate, controls are listed and the selection of consequences by ACO. The route of an ant is regarding consequent actions combination which is selected from every rule of ACO. Among the combination of consequences, the best one is in searching which involves using the heuristic values and pheromone matrix. The FCACO performance is verified on the fuzzy clustering design with a neural network, temperature control on the water bath, and chaotic system control simulations on control on a non-linear system are performed. Simulations on the aforementioned questions and comparisons with other alternative algorithms have influenced FCACO’s performance.