Multivariate Piecewise Rand Divergencive Cockroach Swarm Optimization for Agriculture Crops Recommendation

Abstract

Precision agriculture focuses on monitoring, management information scheme, and variable rate technologies in cropping systems. The primary advantages of precision agriculture results in enhancing crop recommendations, with optimal soil and environmental impact. Crop recommendation, predicting the most suitable crops for a specific region or farm based on factors such as soil type, weather conditions (including temperature, soil pH, and rainfall), is a challenging task in agriculture. Various approaches have been developed for predicting crop recommendations; however, achieving essential factors for timely and accurate detection remains difficult. In this paper, we introduce a novel technique called Multivariate Piecewise Rand Divergencive Cockroach Swarm Optimization (MPRDCSO) for accurate crop recommendation with minimum time consumption. The proposed MPRDCSO technique begins by gathering information from the crop recommendation dataset and consists of two significant steps: data preprocessing and feature selection. In the first step, data preprocessing is carried out to clean and transform input data using Multivariate Piecewise Constant Weighted Interpolation and Camargo's index-based preprocessing for accurate crop recommendation. Following this, the feature extraction process is executed by applying Rand Indexive Jensen–Shannon Divergenced Cockroach Swarm Optimization. In the feature extraction step, the number of features is collected from the dataset, and the rand similarity between the features is measured in the fitness measure to identify the most relevant features. With the extracted optimal attributes, crop recommendation is performed with higher accuracy. An experimental assessment of proposed technique is conducted with accuracy, precision, recall, F1-score, and crop recommendation time across different instances. The quantitatively discussed results indicate that the performance of the proposed MPRDCSO technique achieves higher accuracy with a reduced processing time compared to conventional methods.