Adamax Crop Prediction Neural Network Classification Model Using Iot Data For Green House

Abstract

Crop selection plays an important role in greenhouse cultivation. Greenhouse cultivation is found to be practised by most of the people to meet out their day to day requirements. The objective of this work is to find crop which suits the soil. The above objective is met with the help of various sensors like soil moisture sensor, environment humidity and temperature sensor and pH sensor, apart from this three other values such as N-P-K considered as macro nutrients. Data generated from the above sensors was gathered and pre-processed for further analysis. A neural network based classification model for crop prediction is proposed. The proposed model is compared with standardized algorithm and evaluated with the metrics such as Accuracy, AUC_ROC score, Precision, Recall, F1 Score and support. The obtained result is compared with various optimization algorithms of machine learning. Classification algorithm like Random Forest, SVM, KNN and NAIVE BAYES were used for comparison. The performance of the neural network based classification model seems to be good with an accuracy of 98.6 % and AUC ROC score seems to be 0.991.