Prediction of Plant Growth Through Nutrient Uptake in the Hydroponics System Using Machine Learning Approach

Abstract

AbstractAgriculture has a significant impact on the nation’s economy. Hydroponic gardening is getting more popular as modern agriculture, which allows plants to be grown without ground using liquid fertilizer. To control the growth of hydroponic plants, some work has been performed using machine learning approaches such as neural network models and Bayesian networks. The Internet of things allows machine to machine interface along with artificially intelligent control of the hydroponic system. The proposed study utilizes ML techniques to predict plant damage, nutrient supply per week and crop yield management under a different scenario. A comparison is presented that uses machine learning methods such as logistic regression, decision trees, support vector machine, and random forest to evaluate the performance of the various methods using the mean square error criterion and % accuracy of the model. Various hyper-parameter tuning is executed on a machine learning model to give the best possible accuracy and performance on a given dataset of hydroponic farming. Toward the end, after performing the comparison of different classification algorithms, it is found that support vector machine classifier had performed the best with an average accuracy of 83.67%. So hence by deploying the support vector classifier model, hydroponic farming can be used for better crop management.KeywordsHydroponic farmingCrop yieldSupport vector machineRandom forest