Temperature Prediction Using Recurrent Neural Network for Internet of Things Room Controlling Application
Haider K. Hoomod, Zahraa Sabeeh Amory · 2020
Prediction has several benefits in improving our lives such as in health, weather, seismology, and safety in smart building that able to forecast future values based on past information. Deep learning models for prediction on time series sensor data and choose the appropriate model that fit our case is analyzed. Smart buildings have to make distributed analyze for time series from various sensors in rooms. In this paper, a Recurrent Neural Network (RNM) in particular with the Long Short Term Memory (LSTM) techniques used for advanced analysis of predicting the temperature. The performance of LSTM NN was measured using root mean square error (RMSE) and the results show the model is significantly more accurate.