Key points are not available for this paper at this time.
A crucial development in recent agriculture is the emergence of crop recommendation systems, which provide a unique method for optimizing crop yields and resource use. In this work, we have proposed an advanced "Crop Recommendation System" to help farmers choose the right crop according to different soil and environmental parameters. Our system is trying to improve crop yields and agricultural productivity using machine learning (ML) and deep learning (DL) models. The data used for training and evaluation was collected from the Kaggle dataset repository. It includes the eight most important features such as nitrogen (N), phosphorus (P), potassium (K), rainfall, pH, temperature, humidity, and crop labels representing 22 different crops. This work employs machine learning and deep learning models such as Decision Trees, Random Forests, XGBoost, Support Vector Machine, K-nearest neighbors, Naive Bayes, Artificial Neural Networks, Deep Neural Networks, and Temporal Convolutional Networks. Based on performance evaluation, Random Forest and TCN are the best models, showing 99.2% and 99.9% accuracy, respectively. The performance of other models was also excellent, with accuracy levels ranging from 93.8% to 98.7%. SVM was a little behind, but still performed well with 93.4% accuracy. The project also includes parameter optimization for XGBoost and TCN. The TCN model gives a better result than XGBoost using parameter optimization. The outcome of this research shows that machine learning and deep learning models perform well in crop recommendation systems and the TCN model demonstrates its commitment to accurate and efficient crop recommendations. This research work supports precision agriculture by providing a web interface to the farmers with reliable parameters to improve crop selection based on environment and soil conditions.
Ghosh et al. (Fri,) studied this question.