IOT-BASED GREENHOUSE MONITORING SYSTEM FOR TEMPERATURE AND HUMIDITY PREDICTION AND WEATHER CLASSIFICATION USING LSTM AND SVM MODELS

Although greenhouse cultivation enables year-round crop production in controlled
environments, it requires careful monitoring to maintain ideal conditions. This
study aims to enhance greenhouse management by integrating IoT, ML, and DL
technologies. Data on temperature and humidity, collected by DHT11 sensors
inside and outside the greenhouse, is transmitted via ESP-NOW and stored in a
MySQL database. The LSTM model predicts future temperature and humidity,
while the SVM model classifies weather conditions. The LSTM model showed
superior performance, with RMSE of 0.0766 and MAE of 0.0454 for temperature,
and RMSE of 0.1520 and MAE of 0.1106 for humidity. The SVM model achieved
high classification accuracy with an accuracy of 0.63, precision of 0.64, recall of
0.63, and an F1 score of 0.63, outperforming other models. These results confirm
the effectiveness of the LSTM and SVM models in accurately predicting and
classifying greenhouse conditions, optimizing climate control, reducing energy
consumption, and improving crop yield and quality.
Keywords: Greenhouse, Long Short-Term Memory (LSTM), Support Vector
Machines (SVM), ESP-NOW