Evaluation of users' attention and navigation in a 2D and 3D environment game by using NIPGBoard visualization tool

In the thesis, I present the outlier searching game for two different environments – 2D or computer-based and 3D or Virtual Reality. The experiments were conducted with 30 participants, they were asked to play the game in both environments with different time limits – 10 or 15 minutes. The experiment consisted of 3 main phases 1) instruction period, where the goal of the experiment and game logic was explained, the GDPR form was signed and pen and paper attention test was conducted, 2) training part, where participants learned how to navigate in both environments 3) actual experiment part, where game stages with different setups had been done and different questionnaires (Big Five Inventory, Sense of Presence, User Experience forms and others) had been filled. The goal of the research is to evaluate and compare users’ navigation techniques and performances in two environments - 2D and 3D - while playing a simple outlier searching game. To do that, all the necessary logs had been saved, e. g. mouse/controller clicks, mouse moves, gaze direction vector (in 3D) and gaze screen coordinates (in 2D) etc. I wanted to understand weather there is any connection between user's performance in the game or attention test and his/her personality traits form the BFI-2 form. Two classification tasks had been solved: by having personality traits from the filled BFI-2 form as X, I wanted to predict the performance in the games – class label Y. In two tasks the class label was different – I had multilabel and binary classification problem. In the regression task, I wanted to predict the number of found objects or number of misclicks based on the personality traits. To solve classification tasks, Support Vector Machine with and without feature selection was used. To solve regression tasks, Support Vector Regression was used. Baes on the achieved results of the trained models, we can state that there is no connection between personality traits and game performance. Additionally, overall statistical analysis was carried out on the data, especially for the user performance evaluation, user actions in 2D, gaze points and flying trajectories in 3D environment were visualized, that can be used in future work. Interesting model performance was achieved between attention test and game performance, as it was found out that people who performed good in the game, also performed good in the attention test. Assume that different users performed different navigation techniques. The first stage of describing users was done successfully, I found performance groups. The navigation analysis is an extension of the evaluation on the longer run, as well as the gaze analysis, based on the already developed plots