Immersive Virtual Reality Case Study

Immersive Virtual Reality (IVE) has been implemented as an advanced experimental equipment in diverse research areas. In occupant behavior study, it has also gained reputations since its potentials serve realistic virtual experience to participants and allow researchers to record their responds. Nevertheless, it is still limited due to instrument and side effects lessening the tolerance of participants exposing to the experiment. As a result, conducting IVE experiment usually confronts by small sample size issue causing obstacles to data analysis such as overfitting, high bias and variance, false power of hypothesis test and so forth. To overcome these problems this problem, the authors utilized the potential of Hybrid Hidden Morkov Model (HMM)
The outcome of synthetic data was evaluated and compared with IVE data. The results of cross validation, root mean square errors, and comparison of distribution showed that the hybrid model precisely estimated and generated the synthetic IVE data. (-- removed HTML --) >
INTRODUCTION
Immersive Virtual Reality (IVE) has been implemented in building occupant behavioral studies for many reasons. (-- removed HTML --) >. It allows users to practice and conduct experiments without real environments (in-situ), benefiting studies of occupant behaviors in non-existing or future buildings. It helps the designer to understand behavior of real occupants and optimize the building configurations so that the occupant preferences are properly served. Compared to experiments in in-situ, IVE
In IVE, the tolerance of participants is limited due to instruments and the side effects of IVE such as limited experimental space and time, realness, the inconvenience of using instruments, virtual reality sickness, and so forth. Due to those restrictions, it is expensive in both times and resources to have a large pool of data to accurately analyze outcomes. Furthermore, IVE experiments sometimes are confronted by the issue of small sample size [1], causing obstacles to data analysis processes such as the degree of overfitting[2] [3], high bias and variance [4], false power of hypothesis test, and many more [5] [6] [7] [8]. Moreover, IVE data may be used in higher critical applications rather than basic statistical analysis, e.g. building predictive models which require a large amount of samples to evaluate an accuracy with a low error rate [9]. Hence, the challenges of having sufficient samples are crucial for building occupant behavioral studies using IVE as experimental