Mean Square Error Analysis of Temperature and Humidity in Telkom University Landmark Tower (TULT) Using Multiple Linear Regression (MLR) and Internet of Things (IoT)

Differences in temperature and humidity are observed across various regions and altitudes, attributed to factors such as altitude’s inherent effects like solar radiation and the influence of surrounding winds. Daytime hours typically exhibit higher air humidity levels, a phenomenon similarly observed in building heights. Take, for instance, the Telkom University Landmark Tower (TULT) building. Varied heights within the structure influence temperature and humidity levels. In the academic year 2022/2023, Telkom University in Bandung, West Java, admitted 8,036 new students, a fraction of whom engage in lectures within the TULT building. It is plausible that the influx of students through different entrances to TULT could impact temperature at various altitudes. To address this, the application of Multiple Linear Regression (MLR) and Internet of Things (IoT) methodologies becomes pertinent for predicting temperature and humidity at distinct TULT heights. This involves utilizing the Blynk application for data collection and analysis. Subsequently, the MLR and IoT algorithms are employed to estimate temperature and humidity values. The experimental results are then scrutinized utilizing the Mean Squared Error (MSE) to anticipate error values for each trial. Additionally, the coefficient of determination, R2 or R Square, is utilized for analysis. The MLR model, utilizing training data, exhibits superior performance in predicting temperature and humidity on the 6th floor of the TULT building compared to using testing data. The MLR model effectively elucidates data variability in temperature and humidity when trained data is employed.