Anilchoudary Rugaramji, a computational science and engineering student at the University of Rostock, will defend his master's thesis on "Towards Forecasting Human Energy States: Comparative Analysis and Application of Approaches" on 11.11.2021, at 13:00.
Supervisors and reviewers were Fabienne Lambusch and Prof. Michael Fellmann from the University of Rostock.
The defense will take place in a virtual ZOOM room. All interested parties please register by Thursday, 11.10.2021, 08:00 a.m., to submit dial-in data via email to Fabienne Lambusch (fabienne.lambuschuni-rostockde).
Abstract:
A unifying thread among phenomena such as vitality, vigor, and tiredness is that they all refer to some aspect of energy. Many employees these days are subjected to high work demands, work pressure, and time-constrained expectations, which lead to mental stress and can affect health in the long run. So it is vital to know the energy levels of each individual beforehand to handle these situations well. Forecasting these energy levels may assist individuals in anticipating their energy levels and planning their work accordingly. For example, highly expected creative work can be performed when the energy is high, and a few preventive measures can be taken in advance if we know that the energy predicted is going to be low. Weigelt et al. developed and examined the validity of a single-item pictorial scale of energetic activation. It is proven from various experiments that the scale is valid, time-efficient, and user-friendly. This measure has been used to collect data in social assistance, recovery, and other activities. Understanding individual energies enables employees to assess their strain and recovery, which in turn maintains health and quality of life. The goal of the work is to comprehend these energy curves and forecast human energy states using various statistics and machine learning algorithms and comparing the results across approaches. Throughout the thesis, human energy is an important psychological element. The identical set of classical statistical models is used to examine two univariate datasets. For both datasets, accuracy, mean absolute error, and root mean square error are calculated for all models. In addition, one multivariate data set is analyzed using a combination of statistical and machine learning methods. Finally, a vitality analysis, similar to that of human energy, is performed, and the accuracies are compared. The results indicate that the moving average model outperforms the other models in both the univariate datasets. The SARIMAX offers the highest accuracy for both vitality and battery energy aspects in the multivariate instance.