On the analysis of Amaranthus Viridis crop growth rate

This paper investigates the automation of the crop growth rate, a use case of Amaranthus Viridis. The Linear Regression, Support Vector Regressor (SVR), Decision Tree, K-neighbour, Theil-sen, Quantile Regressor, XGBoost, CatBoost, AdaBoost, Random Forest, Gradient Boosting and Extra Trees Regressor models have been used to compare the crop growth rate. The crop has been evaluated by splitting the data set into training and testing samples. Performance metrics such as Mean Squared Error (MSE), Mean Absolute Error (MAE), Root Mean Square Error (RMSE) and Coefficient of Determinant R² have been used to analyse the dataset. The research has been to automate the prediction of the crop grown within a floating hydroponic system using non-ensemble and ensemble machine learning models. It can be inferred that the XGBoost outperformed the other machine learning models predicting the crop growth rate. The results indicate the Extra Tree regressor produced the highest R², and the SVR produced the lowest R². The XGBoost model predicted the highest crop growth rate and the AdaBoost model predicted the lowest. The highest MSE and MAE have been obtained from the Random Forest Regressor and AdaBoost Regressor. The AdaBoost Regressor has produced the highest RMSE. A comparative analysis of the ensemble and non-ensemble models of the crop growth rate has determined the XGBoost as the optimal model for predicting the crop growth rate in a Floating Hydroponic system.