Enhanced crop recommendation techniques for precision agriculture with deep learning models and optimized feature selection

Abstract

Crop recommendation plays a crucial role in helping farmers make newlineinformed decisions about how to plant based on factors like soil quality, newlineclimate, and market demand. However, some challenges come with crop newlinerecommendations. One challenge is the variability of environmental newlineconditions, which can impact the success of the recommended crops. newlineAdditionally, the availability and accuracy of data, such as historical weather newlinepatterns newlineand soil composition, are essential for making precise newlinerecommendations. Another challenge is the need for continuous monitoring newlineand updating the recommendations to adapt for changing conditions and newlinemarket trends. Overcoming these challenges requires advanced technology, newlinedata analytics, and collaboration between agricultural experts and farmers to newlineensure sustainable and productive crop choices. Thus, the first contributions newlineinvestigate an innovative crop recommendation technique for precise farming newlinewith the goal of increasing crop output while reducing farmer losses. At first, newlinethis study attempts to collect regular data on the agricultural factors of certain newlineplaces. Then, deep features are extracted utilizing an autoencoder, statistical, newlineand PCA-based features are also extracted. The three distinct types of newlinecharacteristics are then combined and sent into the designed AHGSO to newlinedetermine the optimum features. Lastly, the selected optimal characteristics newlineare passed into the recommendation step, wherein a GRU-RC is offered to newlineobtain a precise result about the recommended crop based on the agriculture newlinevariable. The optimum results can be obtained by tuning the parameters newlineassociated with the GRU as well as the ridge classifier using the suggested newlineAHGSO. Finally, the results achieved using the hybrid technique might be newlinedeemed more effective. If a substantial amount of real-time data is gathered to newlinebe processed using the recommended technique, more memory may be newlinerequired for training the algorithm. newline