Monitoring and Prediction of Temperature and Humidity at Telkom University Landmark Tower (TULT) Using Generalized Additive Model (GAM) and Internet of Things (IoT)

Building with many floors requires complex air quality modelling. Telkom University Landmark Tower (TULT) is an impressive 19 floor building, each floor varying in height, which poses challenges in humidity and temperature modelling due to the variations at different altitudes. To address this, the study introduces a solution by employing the Generalized Additive Model (GAM) and Internet of Things (IoT) methods to accurately model humidity and temperature conditions at different heights on TULT. The authors developed a flexible system combining GAM and IoT algorithms for predicting light, humidity, and temperature, enabling precise estimations at various altitudes. The experiments were carried out using Mean Absolute Error (MAE) and Mean Bias Error (MBE) measurements for both test and training data. The results showcased the effectiveness of the approach, with temperature measurements producing a low MAE value of 1.7869 and a negligible MBE value of 0.0506 for the training data. Likewise, humidity measurements revealed an MAE value of 13.3275 and an MBE value of 0.1837 for training data, further confirming the accuracy and reliability of the proposed model. Despite the observed overfitting, it can be noticed that the GAM achieves low error in terms of MBE and MAE. Further hyperparameter optimization is required to prevent the overfitting.